Authors: Ange-Clement Akazan, Verlon Roel Mbingui, Gnankan Landry Regis N'guessan, Issa Karambal
Abstract: Weather forecasting is crucial for managing risks and economic planning, particularly in tropical Africa, where extreme events severely impact livelihoods. Yet, existing forecasting methods often struggle with the region's complex, non-linear weather patterns. This study benchmarks deep recurrent neural networks such as $\texttt{LSTM, GRU, BiLSTM, BiGRU}$, and Kolmogorov-Arnold-based models $(\texttt{KAN} and \texttt{TKAN})$ for daily forecasting of temperature, precipitation, and pressure in two tropical cities: Abidjan, Cote d'Ivoire (Ivory Coast) and Kigali (Rwanda). We further introduce two customized variants of $ \texttt{TKAN}$ that replace its original $\texttt{SiLU}$ activation function with $ \texttt{GeLU}$ and \texttt{MiSH}, respectively. Using station-level meteorological data spanning from 2010 to 2024, we evaluate all the models on standard regression metrics. $\texttt{KAN}$ achieves temperature prediction ($R^2=0.9986$ in Abidjan, $0.9998$ in Kigali, $\texttt{MSE} < 0.0014~^\circ C ^2$), while $\texttt{TKAN}$ variants minimize absolute errors for precipitation forecasting in low-rainfall regimes. The customized $\texttt{TKAN}$ models demonstrate improvements over the standard $\texttt{TKAN}$ across both datasets. Classical \texttt{RNNs} remain highly competitive for atmospheric pressure ($R^2 \approx 0.83{-}0.86$), outperforming $\texttt{KAN}$-based models in this task. These results highlight the potential of spline-based neural architectures for efficient and data-efficient forecasting.
Authors: Jialong Guo, Xinghao Chen, Yehui Tang, Yunhe Wang
Abstract: Large language models(LLMs) have garnered significant attention and demonstrated impressive capabilities in a wide range of applications. However, due to their enormous computational costs, the deployment and application of LLMs are often severely limited. To address this issue, structured pruning is an effective solution to compress the parameters of LLMs. Determining the importance of each sub-module in LLMs and minimizing performance loss are critical issues that need to be carefully addressed in structured pruning. In this paper, we propose an effective and fast structured pruning method named SlimLLM for large language models. For channel and attention head pruning, we evaluate the importance based on the entire channel or head, rather than merely aggregating the importance of individual elements within a sub-module. This approach enables a more holistic consideration of the interdependence among elements within the sub-module. In addition, we design a simple linear regression strategy for the output matrix to quickly recover performance. We also propose layer-based importance ratio to determine the pruning ratio for each layer. Based on the LLaMA benchmark results, our SlimLLM outperforms other methods and achieves state-of-the-art performance.
Authors: Dacao Zhang, Kun Zhang, Shimao Chu, Le Wu, Xin Li, Si Wei
Abstract: With the rapid development of Large Language Models (LLMs), Parameter-Efficient Fine-Tuning (PEFT) methods have gained significant attention, which aims to achieve efficient fine-tuning of LLMs with fewer parameters. As a representative PEFT method, Low-Rank Adaptation (LoRA) introduces low-rank matrices to approximate the incremental tuning parameters and achieves impressive performance over multiple scenarios. After that, plenty of improvements have been proposed for further improvement. However, these methods either focus on single-task scenarios or separately train multiple LoRA modules for multi-task scenarios, limiting the efficiency and effectiveness of LoRA in multi-task scenarios. To better adapt to multi-task fine-tuning, in this paper, we propose a novel Mixture of Low-Rank Experts (MoRE) for multi-task PEFT. Specifically, instead of using an individual LoRA for each task, we align different ranks of LoRA module with different tasks, which we named low-rank experts. Moreover, we design a novel adaptive rank selector to select the appropriate expert for each task. By jointly training low-rank experts, MoRE can enhance the adaptability and efficiency of LoRA in multi-task scenarios. Finally, we conduct extensive experiments over multiple multi-task benchmarks along with different LLMs to verify model performance. Experimental results demonstrate that compared to traditional LoRA and its variants, MoRE significantly improves the performance of LLMs in multi-task scenarios and incurs no additional inference cost. We also release the model and code to facilitate the community.
Authors: Tengfei Lyu, Siyuan Feng, Hao Liu, Hai Yang
Abstract: Ride-hailing platforms face significant challenges in optimizing order dispatching and driver repositioning operations in dynamic urban environments. Traditional approaches based on combinatorial optimization, rule-based heuristics, and reinforcement learning often overlook driver income fairness, interpretability, and adaptability to real-world dynamics. To address these gaps, we propose LLM-ODDR, a novel framework leveraging Large Language Models (LLMs) for joint Order Dispatching and Driver Repositioning (ODDR) in ride-hailing services. LLM-ODDR framework comprises three key components: (1) Multi-objective-guided Order Value Refinement, which evaluates orders by considering multiple objectives to determine their overall value; (2) Fairness-aware Order Dispatching, which balances platform revenue with driver income fairness; and (3) Spatiotemporal Demand-Aware Driver Repositioning, which optimizes idle vehicle placement based on historical patterns and projected supply. We also develop JointDR-GPT, a fine-tuned model optimized for ODDR tasks with domain knowledge. Extensive experiments on real-world datasets from Manhattan taxi operations demonstrate that our framework significantly outperforms traditional methods in terms of effectiveness, adaptability to anomalous conditions, and decision interpretability. To our knowledge, this is the first exploration of LLMs as decision-making agents in ride-hailing ODDR tasks, establishing foundational insights for integrating advanced language models within intelligent transportation systems.
Authors: Eleni Nisioti, Joachim Winther Pedersen, Erwan Plantec, Milton L. Montero, Sebastian Risi
Abstract: The ability to continuously and efficiently transfer skills across tasks is a hallmark of biological intelligence and a long-standing goal in artificial systems. Reinforcement learning (RL), a dominant paradigm for learning in high-dimensional control tasks, is known to suffer from brittleness to task variations and catastrophic forgetting. Neuroevolution (NE) has recently gained attention for its robustness, scalability, and capacity to escape local optima. In this paper, we investigate an understudied dimension of NE: its transfer learning capabilities. To this end, we introduce two benchmarks: a) in stepping gates, neural networks are tasked with emulating logic circuits, with designs that emphasize modular repetition and variation b) ecorobot extends the Brax physics engine with objects such as walls and obstacles and the ability to easily switch between different robotic morphologies. Crucial in both benchmarks is the presence of a curriculum that enables evaluating skill transfer across tasks of increasing complexity. Our empirical analysis shows that NE methods vary in their transfer abilities and frequently outperform RL baselines. Our findings support the potential of NE as a foundation for building more adaptable agents and highlight future challenges for scaling NE to complex, real-world problems.
Authors: Filippo Rinaldi, Giacomo Capitani, Lorenzo Bonicelli, Donato Crisostomi, Federico Bolelli, Elisa Ficarra, Emanuele Rodol\`a, Simone Calderara, Angelo Porrello
Abstract: Foundation models serve as the backbone for numerous specialized models developed through fine-tuning. However, when the underlying pretrained model is updated or retrained (e.g., on larger and more curated datasets), the fine-tuned model becomes obsolete, losing its utility and requiring retraining. This raises the question: is it possible to transfer fine-tuning to a new release of the model? In this work, we investigate how to transfer fine-tuning to a new checkpoint without having to re-train, in a data-free manner. To do so, we draw principles from model re-basin and provide a recipe based on weight permutations to re-base the modifications made to the original base model, often called task vector. In particular, our approach tailors model re-basin for Transformer models, taking into account the challenges of residual connections and multi-head attention layers. Specifically, we propose a two-level method rooted in spectral theory, initially permuting the attention heads and subsequently adjusting parameters within select pairs of heads. Through extensive experiments on visual and textual tasks, we achieve the seamless transfer of fine-tuned knowledge to new pre-trained backbones without relying on a single training step or datapoint. Code is available at https://github.com/aimagelab/TransFusion.
Authors: Mohammad Yaghini, Tudor Cebere, Michael Menart, Aur\'elien Bellet, Nicolas Papernot
Abstract: Many problems in trustworthy ML can be formulated as minimization of the model error under constraints on the prediction rates of the model for suitably-chosen marginals, including most group fairness constraints (demographic parity, equality of odds, etc.). In this work, we study such constrained minimization problems under differential privacy (DP). Standard DP optimization techniques like DP-SGD rely on the loss function's decomposability into per-sample contributions. However, rate constraints introduce inter-sample dependencies, violating the decomposability requirement. To address this, we develop RaCO-DP, a DP variant of the Stochastic Gradient Descent-Ascent (SGDA) algorithm which solves the Lagrangian formulation of rate constraint problems. We demonstrate that the additional privacy cost of incorporating these constraints reduces to privately estimating a histogram over the mini-batch at each optimization step. We prove the convergence of our algorithm through a novel analysis of SGDA that leverages the linear structure of the dual parameter. Finally, empirical results on learning under group fairness constraints demonstrate that our method Pareto-dominates existing private learning approaches in fairness-utility trade-offs.
Authors: Aniruddha Nrusimha, William Brandon, Mayank Mishra, Yikang Shen, Rameswar Panda, Jonathan Ragan-Kelley, Yoon Kim
Abstract: The size and compute characteristics of modern large language models have led to an increased interest in developing specialized kernels tailored for training and inference. Existing kernels primarily optimize for compute utilization, targeting the large-batch training and inference settings. However, low-batch inference, where memory bandwidth and kernel launch overheads contribute are significant factors, remains important for many applications of interest such as in edge deployment and latency-sensitive applications. This paper describes FlashFormer, a proof-of-concept kernel for accelerating single-batch inference for transformer-based large language models. Across various model sizes and quantizations settings, we observe nontrivial speedups compared to existing state-of-the-art inference kernels.
Authors: Divya Shanmugam, Helen Lu, Swami Sankaranarayanan, John Guttag
Abstract: A conformal classifier produces a set of predicted classes and provides a probabilistic guarantee that the set includes the true class. Unfortunately, it is often the case that conformal classifiers produce uninformatively large sets. In this work, we show that test-time augmentation (TTA)--a technique that introduces inductive biases during inference--reduces the size of the sets produced by conformal classifiers. Our approach is flexible, computationally efficient, and effective. It can be combined with any conformal score, requires no model retraining, and reduces prediction set sizes by 10%-14% on average. We conduct an evaluation of the approach spanning three datasets, three models, two established conformal scoring methods, different guarantee strengths, and several distribution shifts to show when and why test-time augmentation is a useful addition to the conformal pipeline.
Authors: Mert Onur Cakiroglu, Idil Bilge Altun, Hasan Kurban, Elham Buxton, Mehmet Dalkilic
Abstract: Time series forecasting remains a challenging task for foundation models due to temporal heterogeneity, high dimensionality, and the lack of inherent symbolic structure. In this work, we propose DRAGON (Discrete Representation and Augmented Graph encoding Over deBruijN Graphs), a novel encoder that introduces Multivariate de Bruijn Graphs (MdBGs) to bridge the gap between symbolic representations and neural modeling. DRAGON discretizes continuous input sequences and maps them onto a fixed graph structure, enabling dynamic context recovery via graph-based attention. Integrated as an auxiliary module within a dual-branch architecture, DRAGON augments conventional CNN-based encoders with symbolic, structure-aware representations. All code developed for this study is available at: https://github.com/KurbanIntelligenceLab/MultdBG-Time-Series-Library
URLs: https://github.com/KurbanIntelligenceLab/MultdBG-Time-Series-Library
Authors: Claas Voelcker, Anastasiia Pedan, Arash Ahmadian, Romina Abachi, Igor Gilitschenski, Amir-massoud Farahmand
Abstract: The idea of value-aware model learning, that models should produce accurate value estimates, has gained prominence in model-based reinforcement learning. The MuZero loss, which penalizes a model's value function prediction compared to the ground-truth value function, has been utilized in several prominent empirical works in the literature. However, theoretical investigation into its strengths and weaknesses is limited. In this paper, we analyze the family of value-aware model learning losses, which includes the popular MuZero loss. We show that these losses, as normally used, are uncalibrated surrogate losses, which means that they do not always recover the correct model and value function. Building on this insight, we propose corrections to solve this issue. Furthermore, we investigate the interplay between the loss calibration, latent model architectures, and auxiliary losses that are commonly employed when training MuZero-style agents. We show that while deterministic models can be sufficient to predict accurate values, learning calibrated stochastic models is still advantageous.
Authors: Sarah Meiklejohn, Hayden Blauzvern, Mihai Maruseac, Spencer Schrock, Laurent Simon, Ilia Shumailov
Abstract: Powerful machine learning (ML) models are now readily available online, which creates exciting possibilities for users who lack the deep technical expertise or substantial computing resources needed to develop them. On the other hand, this type of open ecosystem comes with many risks. In this paper, we argue that the current ecosystem for open ML models contains significant supply-chain risks, some of which have been exploited already in real attacks. These include an attacker replacing a model with something malicious (e.g., malware), or a model being trained using a vulnerable version of a framework or on restricted or poisoned data. We then explore how Sigstore, a solution designed to bring transparency to open-source software supply chains, can be used to bring transparency to open ML models, in terms of enabling model publishers to sign their models and prove properties about the datasets they use.
Authors: Marco Fumero, Luca Moschella, Emanuele Rodol\`a, Francesco Locatello
Abstract: Neural networks transform high-dimensional data into compact, structured representations, often modeled as elements of a lower dimensional latent space. In this paper, we present an alternative interpretation of neural models as dynamical systems acting on the latent manifold. Specifically, we show that autoencoder models implicitly define a latent vector field on the manifold, derived by iteratively applying the encoding-decoding map, without any additional training. We observe that standard training procedures introduce inductive biases that lead to the emergence of attractor points within this vector field. Drawing on this insight, we propose to leverage the vector field as a representation for the network, providing a novel tool to analyze the properties of the model and the data. This representation enables to: (i) analyze the generalization and memorization regimes of neural models, even throughout training; (ii) extract prior knowledge encoded in the network's parameters from the attractors, without requiring any input data; (iii) identify out-of-distribution samples from their trajectories in the vector field. We further validate our approach on vision foundation models, showcasing the applicability and effectiveness of our method in real-world scenarios.
Authors: Anton Bj\"orklund, Mykola Zaitsev, Marta Kwiatkowska
Abstract: The growing reliance on artificial intelligence in safety- and security-critical applications demands effective neural network certification. A challenging real-world use case is certification against ``patch attacks'', where adversarial patches or lighting conditions obscure parts of images, for example traffic signs. One approach to certification, which also gives quantitative coverage estimates, utilizes preimages of neural networks, i.e., the set of inputs that lead to a specified output. However, these preimage approximation methods, including the state-of-the-art PREMAP algorithm, struggle with scalability. This paper presents novel algorithmic improvements to PREMAP involving tighter bounds, adaptive Monte Carlo sampling, and improved branching heuristics. We demonstrate efficiency improvements of at least an order of magnitude on reinforcement learning control benchmarks, and show that our method scales to convolutional neural networks that were previously infeasible. Our results demonstrate the potential of preimage approximation methodology for reliability and robustness certification.
Authors: Pedro Mendes, Paolo Romano, David Garlan
Abstract: Reliable uncertainty estimation is critical for deploying neural networks (NNs) in real-world applications. While existing calibration techniques often rely on post-hoc adjustments or coarse-grained binning methods, they remain limited in scalability, differentiability, and generalization across domains. In this work, we introduce CLUE (Calibration via Learning Uncertainty-Error Alignment), a novel approach that explicitly aligns predicted uncertainty with observed error during training, grounded in the principle that well-calibrated models should produce uncertainty estimates that match their empirical loss. CLUE adopts a novel loss function that jointly optimizes predictive performance and calibration, using summary statistics of uncertainty and loss as proxies. The proposed method is fully differentiable, domain-agnostic, and compatible with standard training pipelines. Through extensive experiments on vision, regression, and language modeling tasks, including out-of-distribution and domain-shift scenarios, we demonstrate that CLUE achieves superior calibration quality and competitive predictive performance with respect to state-of-the-art approaches without imposing significant computational overhead.
Authors: Josiah Couch, Miao Li, Rima Arnaout, Ramy Arnaout
Abstract: In the universal quest to optimize machine-learning classifiers, three factors -- model architecture, dataset size, and class balance -- have been shown to influence test-time performance but do not fully account for it. Previously, evidence was presented for an additional factor that can be referred to as dataset quality, but it was unclear whether this was actually a joint property of the dataset and the model architecture, or an intrinsic property of the dataset itself. If quality is truly dataset-intrinsic and independent of model architecture, dataset size, and class balance, then the same datasets should perform better (or worse) regardless of these other factors. To test this hypothesis, here we create thousands of datasets, each controlled for size and class balance, and use them to train classifiers with a wide range of architectures, from random forests and support-vector machines to deep networks. We find that classifier performance correlates strongly by subset across architectures ($R^2=0.79$), supporting quality as an intrinsic property of datasets independent of dataset size and class balance and of model architecture. Digging deeper, we find that dataset quality appears to be an emergent property of something more fundamental: the quality of datasets' constituent classes. Thus, quality joins size, class balance, and model architecture as an independent correlate of performance and a separate target for optimizing machine-learning-based classification.
Authors: Ayush Sawarni, Sahasrajit Sarmasarkar, Vasilis Syrgkanis
Abstract: This paper investigates the integration of response time data into human preference learning frameworks for more effective reward model elicitation. While binary preference data has become fundamental in fine-tuning foundation models, generative AI systems, and other large-scale models, the valuable temporal information inherent in user decision-making remains largely unexploited. We propose novel methodologies to incorporate response time information alongside binary choice data, leveraging the Evidence Accumulation Drift Diffusion (EZ) model, under which response time is informative of the preference strength. We develop Neyman-orthogonal loss functions that achieve oracle convergence rates for reward model learning, matching the theoretical optimal rates that would be attained if the expected response times for each query were known a priori. Our theoretical analysis demonstrates that for linear reward functions, conventional preference learning suffers from error rates that scale exponentially with reward magnitude. In contrast, our response time-augmented approach reduces this to polynomial scaling, representing a significant improvement in sample efficiency. We extend these guarantees to non-parametric reward function spaces, establishing convergence properties for more complex, realistic reward models. Our extensive experiments validate our theoretical findings in the context of preference learning over images.
Authors: Michael Klamkin, Mathieu Tanneau, Pascal Van Hentenryck
Abstract: Machine Learning (ML) techniques for Optimal Power Flow (OPF) problems have recently garnered significant attention, reflecting a broader trend of leveraging ML to approximate and/or accelerate the resolution of complex optimization problems. These developments are necessitated by the increased volatility and scale in energy production for modern and future grids. However, progress in ML for OPF is hindered by the lack of standardized datasets and evaluation metrics, from generating and solving OPF instances, to training and benchmarking machine learning models. To address this challenge, this paper introduces PGLearn, a comprehensive suite of standardized datasets and evaluation tools for ML and OPF. PGLearn provides datasets that are representative of real-life operating conditions, by explicitly capturing both global and local variability in the data generation, and by, for the first time, including time series data for several large-scale systems. In addition, it supports multiple OPF formulations, including AC, DC, and second-order cone formulations. Standardized datasets are made publicly available to democratize access to this field, reduce the burden of data generation, and enable the fair comparison of various methodologies. PGLearn also includes a robust toolkit for training, evaluating, and benchmarking machine learning models for OPF, with the goal of standardizing performance evaluation across the field. By promoting open, standardized datasets and evaluation metrics, PGLearn aims at democratizing and accelerating research and innovation in machine learning applications for optimal power flow problems. Datasets are available for download at https://www.huggingface.co/PGLearn.
Authors: Chenruo Liu, Kenan Tang, Yao Qin, Qi Lei
Abstract: This paper bridges distribution shift and AI safety through a comprehensive analysis of their conceptual and methodological synergies. While prior discussions often focus on narrow cases or informal analogies, we establish two types connections between specific causes of distribution shift and fine-grained AI safety issues: (1) methods addressing a specific shift type can help achieve corresponding safety goals, or (2) certain shifts and safety issues can be formally reduced to each other, enabling mutual adaptation of their methods. Our findings provide a unified perspective that encourages fundamental integration between distribution shift and AI safety research.
Authors: Liu Yuezhang, Xue-Xin Wei
Abstract: Recent findings suggest that diffusion models significantly enhance empirical adversarial robustness. While some intuitive explanations have been proposed, the precise mechanisms underlying these improvements remain unclear. In this work, we systematically investigate how and how well diffusion models improve adversarial robustness. First, we observe that diffusion models intriguingly increase, rather than decrease, the $\ell_p$ distance to clean samples--challenging the intuition that purification denoises inputs closer to the original data. Second, we find that the purified images are heavily influenced by the internal randomness of diffusion models, where a compression effect arises within each randomness configuration. Motivated by this observation, we evaluate robustness under fixed randomness and find that the improvement drops to approximately 24% on CIFAR-10--substantially lower than prior reports approaching 70%. Importantly, we show that this remaining robustness gain strongly correlates with the model's ability to compress the input space, revealing the compression rate as a reliable robustness indicator without requiring gradient-based analysis. Our findings provide novel insights into the mechanisms underlying diffusion-based purification, and offer guidance for developing more effective and principled adversarial purification systems.
Authors: Dharambir Mahto, Prashant Yadav, Mahesh Banavar, Jim Keany, Alan T Joseph, Srinivas Kilambi
Abstract: Sepsis is a life-threatening condition affecting over 48.9 million people globally and causing 11 million deaths annually. Despite medical advancements, predicting sepsis remains a challenge due to non-specific symptoms and complex pathophysiology. The SXI++ LNM is a machine learning scoring system that refines sepsis prediction by leveraging multiple algorithms and deep neural networks. This study aims to improve robustness in clinical applications and evaluates the predictive performance of the SXI++ LNM for sepsis prediction. The model, utilizing a deep neural network, was trained and tested using multiple scenarios with different dataset distributions. The model's performance was assessed against unseen test data, and accuracy, precision, and area under the curve (AUC) were calculated. THE SXI++ LNM outperformed the state of the art in three use cases, achieving an AUC of 0.99 (95% CI: 0.98-1.00). The model demonstrated a precision of 99.9% (95% CI: 99.8-100.0) and an accuracy of 99.99% (95% CI: 99.98-100.0), maintaining high reliability.
Authors: Franck Gabriel, Fran\c{c}ois Ged, Maria Han Veiga, Emmanuel Schertzer
Abstract: Diffusion models now set the benchmark in high-fidelity generative sampling, yet they can, in principle, be prone to memorization. In this case, their learned score overfits the finite dataset so that the reverse-time SDE samples are mostly training points. In this paper, we interpret the empirical score as a noisy version of the true score and show that its covariance matrix is asymptotically a re-weighted data PCA. In large dimension, the small time limit makes the noise variance blow up while simultaneously reducing spatial correlation. To reduce this variance, we introduce a kernel-smoothed empirical score and analyze its bias-variance trade-off. We derive asymptotic bounds on the Kullback-Leibler divergence between the true distribution and the one generated by the modified reverse SDE. Regularization on the score has the same effect as increasing the size of the training dataset, and thus helps prevent memorization. A spectral decomposition of the forward diffusion suggests better variance control under some regularity conditions of the true data distribution. Reverse diffusion with kernel-smoothed empirical score can be reformulated as a gradient descent drifted toward a Log-Exponential Double-Kernel Density Estimator (LED-KDE). This perspective highlights two regularization mechanisms taking place in denoising diffusions: an initial Gaussian kernel first diffuses mass isotropically in the ambient space, while a second kernel applied in score space concentrates and spreads that mass along the data manifold. Hence, even a straightforward regularization-without any learning-already mitigates memorization and enhances generalization. Numerically, we illustrate our results with several experiments on synthetic and MNIST datasets.
Authors: Nikita Khramov, Andrei Kozyrev, Gleb Solovev, Anton Podkopaev
Abstract: Interactive Theorem Proving was repeatedly shown to be fruitful combined with Generative Artificial Intelligence. This paper assesses multiple approaches to Rocq generation and illuminates potential avenues for improvement. We highlight the importance of thorough premise selection for generating Rocq proofs and propose a novel approach, leveraging retrieval via a self-attentive embedder model. The evaluation of the designed approach shows up to 28% relative increase of the generator's performance. We tackle the problem of writing Rocq proofs using a multi-stage agentic system, tailored for formal verification, and demonstrate its high effectiveness. We conduct an ablation study and show the use of multi-agent debate on the planning stage of proof synthesis.
Authors: Carlota Par\'es-Morlans, Michelle Yi, Claire Chen, Sarah A. Wu, Rika Antonova, Tobias Gerstenberg, Jeannette Bohg
Abstract: Tasks that involve complex interactions between objects with unknown dynamics make planning before execution difficult. These tasks require agents to iteratively improve their actions after actively exploring causes and effects in the environment. For these type of tasks, we propose Causal-PIK, a method that leverages Bayesian optimization to reason about causal interactions via a Physics-Informed Kernel to help guide efficient search for the best next action. Experimental results on Virtual Tools and PHYRE physical reasoning benchmarks show that Causal-PIK outperforms state-of-the-art results, requiring fewer actions to reach the goal. We also compare Causal-PIK to human studies, including results from a new user study we conducted on the PHYRE benchmark. We find that Causal-PIK remains competitive on tasks that are very challenging, even for human problem-solvers.
Authors: Nicolas Espinosa-Dice, Yiyi Zhang, Yiding Chen, Bradley Guo, Owen Oertell, Gokul Swamy, Kiante Brantley, Wen Sun
Abstract: Diffusion and flow models have emerged as powerful generative approaches capable of modeling diverse and multimodal behavior. However, applying these models to offline reinforcement learning (RL) remains challenging due to the iterative nature of their noise sampling processes, making policy optimization difficult. In this paper, we introduce Scalable Offline Reinforcement Learning (SORL), a new offline RL algorithm that leverages shortcut models - a novel class of generative models - to scale both training and inference. SORL's policy can capture complex data distributions and can be trained simply and efficiently in a one-stage training procedure. At test time, SORL introduces both sequential and parallel inference scaling by using the learned Q-function as a verifier. We demonstrate that SORL achieves strong performance across a range of offline RL tasks and exhibits positive scaling behavior with increased test-time compute. We release the code at nico-espinosadice.github.io/projects/sorl.
Authors: Hyungki Im, Wyame Benslimane, Paul Grigas
Abstract: We study an extension of contextual stochastic linear optimization (CSLO) that, in contrast to most of the existing literature, involves inequality constraints that depend on uncertain parameters predicted by a machine learning model. To handle the constraint uncertainty, we use contextual uncertainty sets constructed via methods like conformal prediction. Given a contextual uncertainty set method, we introduce the "Smart Predict-then-Optimize with Robust Constraints" (SPO-RC) loss, a feasibility-sensitive adaptation of the SPO loss that measures decision error of predicted objective parameters. We also introduce a convex surrogate, SPO-RC+, and prove Fisher consistency with SPO-RC. To enhance performance, we train on truncated datasets where true constraint parameters lie within the uncertainty sets, and we correct the induced sample selection bias using importance reweighting techniques. Through experiments on fractional knapsack and alloy production problem instances, we demonstrate that SPO-RC+ effectively handles uncertainty in constraints and that combining truncation with importance reweighting can further improve performance.
Authors: Naram Mhaisen, George Iosifidis
Abstract: We revisit the Follow the Regularized Leader (FTRL) framework for Online Convex Optimization (OCO) over compact sets, focusing on achieving dynamic regret guarantees. Prior work has highlighted the framework's limitations in dynamic environments due to its tendency to produce "lazy" iterates. However, building on insights showing FTRL's ability to produce "agile" iterates, we show that it can indeed recover known dynamic regret bounds through optimistic composition of future costs and careful linearization of past costs, which can lead to pruning some of them. This new analysis of FTRL against dynamic comparators yields a principled way to interpolate between greedy and agile updates and offers several benefits, including refined control over regret terms, optimism without cyclic dependence, and the application of minimal recursive regularization akin to AdaFTRL. More broadly, we show that it is not the lazy projection style of FTRL that hinders (optimistic) dynamic regret, but the decoupling of the algorithm's state (linearized history) from its iterates, allowing the state to grow arbitrarily. Instead, pruning synchronizes these two when necessary.
Authors: Youngsoo Choi, Siu Wun Cheung, Youngkyu Kim, Ping-Hsuan Tsai, Alejandro N. Diaz, Ivan Zanardi, Seung Whan Chung, Dylan Matthew Copeland, Coleman Kendrick, William Anderson, Traian Iliescu, Matthias Heinkenschloss
Abstract: The widespread success of foundation models in natural language processing and computer vision has inspired researchers to extend the concept to scientific machine learning and computational science. However, this position paper argues that as the term "foundation model" is an evolving concept, its application in computational science is increasingly used without a universally accepted definition, potentially creating confusion and diluting its precise scientific meaning. In this paper, we address this gap by proposing a formal definition of foundation models in computational science, grounded in the core values of generality, reusability, and scalability. We articulate a set of essential and desirable characteristics that such models must exhibit, drawing parallels with traditional foundational methods, like the finite element and finite volume methods. Furthermore, we introduce the Data-Driven Finite Element Method (DD-FEM), a framework that fuses the modular structure of classical FEM with the representational power of data-driven learning. We demonstrate how DD-FEM addresses many of the key challenges in realizing foundation models for computational science, including scalability, adaptability, and physics consistency. By bridging traditional numerical methods with modern AI paradigms, this work provides a rigorous foundation for evaluating and developing novel approaches toward future foundation models in computational science.
Authors: Donghyeon Joo, Helya Hosseini, Ramyad Hadidi, Bahar Asgari
Abstract: We demonstrate that unstructured sparsity significantly improves KV cache compression for LLMs, enabling sparsity levels up to 70% without compromising accuracy or requiring fine-tuning. We conduct a systematic exploration of pruning strategies and find per-token magnitude-based pruning as highly effective for both Key and Value caches under unstructured sparsity, surpassing prior structured pruning schemes. The Key cache benefits from prominent outlier elements, while the Value cache surprisingly benefits from a simple magnitude-based pruning despite its uniform distribution. KV cache size is the major bottleneck in decode performance due to high memory overhead for large context lengths. To address this, we use a bitmap-based sparse format and a custom attention kernel capable of compressing and directly computing over compressed caches pruned to arbitrary sparsity patterns, significantly accelerating memory-bound operations in decode computations and thereby compensating for the overhead of runtime pruning and compression. Our custom attention kernel coupled with the bitmap-based format delivers substantial compression of KV cache upto 45% of dense inference and thereby enables longer context length and increased tokens/sec throughput of upto 2.23x compared to dense inference. Our pruning mechanism and sparse attention kernel is available at https://github.com/dhjoo98/mustafar.
Authors: Athanasios Glentis, Jiaxiang Li, Qiulin Shang, Andi Han, Ioannis Tsaknakis, Quan Wei, Mingyi Hong
Abstract: Fueled by their remarkable ability to tackle diverse tasks across multiple domains, large language models (LLMs) have grown at an unprecedented rate, with some recent models containing trillions of parameters. This growth is accompanied by substantial computational challenges, particularly regarding the memory and compute resources required for training and fine-tuning. Numerous approaches have been explored to address these issues, such as LoRA. While these methods are effective for fine-tuning, their application to pre-training is significantly more challenging due to the need to learn vast datasets. Motivated by this issue, we aim to address the following questions: Can parameter- or memory-efficient methods enhance pre-training efficiency while achieving performance comparable to full-model training? How can the performance gap be narrowed? To this end, the contributions of this work are the following. (1) We begin by conducting a comprehensive survey that summarizes state-of-the-art methods for efficient pre-training. (2) We perform a benchmark evaluation of several representative memory efficient pre-training approaches to comprehensively evaluate their performance across model sizes. We observe that with a proper choice of optimizer and hyperparameters, full-rank training delivers the best performance, as expected. We also notice that incorporating high-rank updates in low-rank approaches is the key to improving their performance. (3) Finally, we propose two practical techniques, namely weight refactorization and momentum reset, to enhance the performance of efficient pre-training methods. We observe that applying these techniques to the low-rank method (on a 1B model) can achieve a lower perplexity than popular memory efficient algorithms such as GaLore and Fira, while simultaneously using about 25% less memory.
Authors: Jipeng Li, Yanning Shen
Abstract: Explicit noise-level conditioning is widely regarded as essential for the effective operation of Graph Diffusion Models (GDMs). In this work, we challenge this assumption by investigating whether denoisers can implicitly infer noise levels directly from corrupted graph structures, potentially eliminating the need for explicit noise conditioning. To this end, we develop a theoretical framework centered on Bernoulli edge-flip corruptions and extend it to encompass more complex scenarios involving coupled structure-attribute noise. Extensive empirical evaluations on both synthetic and real-world graph datasets, using models such as GDSS and DiGress, provide strong support for our theoretical findings. Notably, unconditional GDMs achieve performance comparable or superior to their conditioned counterparts, while also offering reductions in parameters (4-6%) and computation time (8-10%). Our results suggest that the high-dimensional nature of graph data itself often encodes sufficient information for the denoising process, opening avenues for simpler, more efficient GDM architectures.
Authors: Michael Sun, Orion Foo, Gang Liu, Wojciech Matusik, Jie Chen
Abstract: Directed acyclic graphs (DAGs) are a class of graphs commonly used in practice, with examples that include electronic circuits, Bayesian networks, and neural architectures. While many effective encoders exist for DAGs, it remains challenging to decode them in a principled manner, because the nodes of a DAG can have many different topological orders. In this work, we propose a grammar-based approach to constructing a principled, compact and equivalent sequential representation of a DAG. Specifically, we view a graph as derivations over an unambiguous grammar, where the DAG corresponds to a unique sequence of production rules. Equivalently, the procedure to construct such a description can be viewed as a lossless compression of the data. Such a representation has many uses, including building a generative model for graph generation, learning a latent space for property prediction, and leveraging the sequence representational continuity for Bayesian Optimization over structured data. Code is available at https://github.com/shiningsunnyday/induction.
Authors: Chengqi Zheng, Jianda Chen, Yueming Lyu, Wen Zheng Terence Ng, Haopeng Zhang, Yew-Soon Ong, Ivor Tsang, Haiyan Yin
Abstract: Despite the promise of autonomous agentic reasoning, existing workflow generation methods frequently produce fragile, unexecutable plans due to unconstrained LLM-driven construction. We introduce MermaidFlow, a framework that redefines the agentic search space through safety-constrained graph evolution. At its core, MermaidFlow represent workflows as a verifiable intermediate representation using Mermaid, a structured and human-interpretable graph language. We formulate domain-aware evolutionary operators, i.e., crossover, mutation, insertion, and deletion, to preserve semantic correctness while promoting structural diversity, enabling efficient exploration of a high-quality, statically verifiable workflow space. Without modifying task settings or evaluation protocols, MermaidFlow achieves consistent improvements in success rates and faster convergence to executable plans on the agent reasoning benchmark. The experimental results demonstrate that safety-constrained graph evolution offers a scalable, modular foundation for robust and interpretable agentic reasoning systems.
Authors: Bahareh Tolooshams, Aditi Chandrashekar, Rayhan Zirvi, Abbas Mammadov, Jiachen Yao, Chuwei Wang, Anima Anandkumar
Abstract: Diffusion models represent the state-of-the-art for solving inverse problems such as image restoration tasks. In the Bayesian framework, diffusion-based inverse solvers incorporate a likelihood term to guide the prior sampling process, generating data consistent with the posterior distribution. However, due to the intractability of the likelihood term, many current methods rely on isotropic Gaussian approximations, which lead to deviations from the data manifold and result in inconsistent, unstable reconstructions. We propose Equivariance Regularized (EquiReg) diffusion, a general framework for regularizing posterior sampling in diffusion-based inverse problem solvers. EquiReg enhances reconstructions by reweighting diffusion trajectories and penalizing those that deviate from the data manifold. We define a new distribution-dependent equivariance error, empirically identify functions that exhibit low error for on-manifold samples and higher error for off-manifold samples, and leverage these functions to regularize the diffusion sampling process. When applied to a variety of solvers, EquiReg outperforms state-of-the-art diffusion models in both linear and nonlinear image restoration tasks, as well as in reconstructing partial differential equations.
Authors: Guancheng Zhou, Haiping Xu, Hongkang Xu, Chenyu Li, Donghui Yan
Abstract: The popular K-means clustering algorithm potentially suffers from a major weakness for further analysis or interpretation. Some cluster may have disproportionately more (or fewer) points from one of the subpopulations in terms of some sensitive variable, e.g., gender or race. Such a fairness issue may cause bias and unexpected social consequences. This work attempts to improve the fairness of K-means clustering with a two-stage optimization formulation--clustering first and then adjust cluster membership of a small subset of selected data points. Two computationally efficient algorithms are proposed in identifying those data points that are expensive for fairness, with one focusing on nearest data points outside of a cluster and the other on highly 'mixed' data points. Experiments on benchmark datasets show substantial improvement on fairness with a minimal impact to clustering quality. The proposed algorithms can be easily extended to a broad class of clustering algorithms or fairness metrics.
Authors: Masaharu Kagiyama, Tsuyoshi Okita
Abstract: This paper aims to develop an energy-efficient classifier for time-series data by introducing PatchEchoClassifier, a novel model that leverages a reservoir-based mechanism known as the Echo State Network (ESN). The model is designed for human activity recognition (HAR) using one-dimensional sensor signals and incorporates a tokenizer to extract patch-level representations. To train the model efficiently, we propose a knowledge distillation framework that transfers knowledge from a high-capacity MLP-Mixer teacher to the lightweight reservoir-based student model. Experimental evaluations on multiple HAR datasets demonstrate that our model achieves over 80 percent accuracy while significantly reducing computational cost. Notably, PatchEchoClassifier requires only about one-sixth of the floating point operations (FLOPS) compared to DeepConvLSTM, a widely used convolutional baseline. These results suggest that PatchEchoClassifier is a promising solution for real-time and energy-efficient human activity recognition in edge computing environments.
Authors: Albert Tseng, Zhaofeng Sun, Christopher De Sa
Abstract: The main goal of post-training quantization (PTQ) is to produced a compressed model whose output distribution is as close to the original model's as possible. To do this tractably, almost all LLM PTQ algorithms quantize linear layers by independently minimizing the immediate activation error. However, this localized objective ignores the effect of subsequent layers, so reducing it does not necessarily give a closer model. In this work, we introduce Yet Another Quantization Algorithm (YAQA), an adaptive rounding algorithm that uses Kronecker-factored approximations of each linear layer's Hessian with respect to the \textit{full model} KL divergence. YAQA consists of two components: Kronecker-factored sketches of the full layerwise Hessian that can be tractably computed for hundred-billion parameter LLMs, and a quantizer-independent rounding algorithm that uses these sketches and comes with theoretical guarantees. Across a wide range of models and quantizers, YAQA empirically reduces the KL divergence to the original model by $\approx 30\%$ while achieving state of the art performance on downstream tasks.
Authors: Behzad Kamgar-Parsi, Behrooz Kamgar-Parsi
Abstract: Finding the number of meaningful clusters in an unlabeled dataset is important in many applications. Regularized k-means algorithm is a possible approach frequently used to find the correct number of distinct clusters in datasets. The most common formulation of the regularization function is the additive linear term $\lambda k$, where $k$ is the number of clusters and $\lambda$ a positive coefficient. Currently, there are no principled guidelines for setting a value for the critical hyperparameter $\lambda$. In this paper, we derive rigorous bounds for $\lambda$ assuming clusters are {\em ideal}. Ideal clusters (defined as $d$-dimensional spheres with identical radii) are close proxies for k-means clusters ($d$-dimensional spherically symmetric distributions with identical standard deviations). Experiments show that the k-means algorithm with additive regularizer often yields multiple solutions. Thus, we also analyze k-means algorithm with multiplicative regularizer. The consensus among k-means solutions with additive and multiplicative regularizations reduces the ambiguity of multiple solutions in certain cases. We also present selected experiments that demonstrate performance of the regularized k-means algorithms as clusters deviate from the ideal assumption.
Authors: Ari S. Benjamin, Kyle Daruwalla, Christian Pehle, Anthony M. Zador
Abstract: One frequently wishes to learn a range of similar tasks as efficiently as possible, re-using knowledge across tasks. In artificial neural networks, this is typically accomplished by conditioning a network upon task context by injecting context as input. Brains have a different strategy: the parameters themselves are modulated as a function of various neuromodulators such as serotonin. Here, we take inspiration from neuromodulation and propose to learn weights which are smoothly parameterized functions of task context variables. Rather than optimize a weight vector, i.e. a single point in weight space, we optimize a smooth manifold in weight space with a predefined topology. To accomplish this, we derive a formal treatment of optimization of manifolds as the minimization of a loss functional subject to a constraint on volumetric movement, analogous to gradient descent. During inference, conditioning selects a single point on this manifold which serves as the effective weight matrix for a particular sub-task. This strategy for conditioning has two main advantages. First, the topology of the manifold (whether a line, circle, or torus) is a convenient lever for inductive biases about the relationship between tasks. Second, learning in one state smoothly affects the entire manifold, encouraging generalization across states. To verify this, we train manifolds with several topologies, including straight lines in weight space (for conditioning on e.g. noise level in input data) and ellipses (for rotated images). Despite their simplicity, these parameterizations outperform conditioning identical networks by input concatenation and better generalize to out-of-distribution samples. These results suggest that modulating weights over low-dimensional manifolds offers a principled and effective alternative to traditional conditioning.
Authors: Jihwan Oh, Murad Aghazada, Se-Young Yun, Taehyeon Kim
Abstract: Bargaining, a critical aspect of real-world interactions, presents challenges for large language models (LLMs) due to limitations in strategic depth and adaptation to complex human factors. Existing benchmarks often fail to capture this real-world complexity. To address this and enhance LLM capabilities in realistic bargaining, we introduce a comprehensive framework centered on utility-based feedback. Our contributions are threefold: (1) BargainArena, a novel benchmark dataset with six intricate scenarios (e.g., deceptive practices, monopolies) to facilitate diverse strategy modeling; (2) human-aligned, economically-grounded evaluation metrics inspired by utility theory, incorporating agent utility and negotiation power, which implicitly reflect and promote opponent-aware reasoning (OAR); and (3) a structured feedback mechanism enabling LLMs to iteratively refine their bargaining strategies. This mechanism can positively collaborate with in-context learning (ICL) prompts, including those explicitly designed to foster OAR. Experimental results show that LLMs often exhibit negotiation strategies misaligned with human preferences, and that our structured feedback mechanism significantly improves their performance, yielding deeper strategic and opponent-aware reasoning.
Authors: Linh Le Pham Van, Minh Hoang Nguyen, Hung Le, Hung The Tran, Sunil Gupta
Abstract: Robust reinforcement learning (RL) aims to learn policies that remain effective despite uncertainties in its environment, which frequently arise in real-world applications due to variations in environment dynamics. The robust RL methods learn a robust policy by maximizing value under the worst-case models within a predefined uncertainty set. Offline robust RL algorithms are particularly promising in scenarios where only a fixed dataset is available and new data cannot be collected. However, these approaches often require extensive offline data, and gathering such datasets for specific tasks in specific environments can be both costly and time-consuming. Using an imperfect simulator offers a faster, cheaper, and safer way to collect data for training, but it can suffer from dynamics mismatch. In this paper, we introduce HYDRO, the first Hybrid Cross-Domain Robust RL framework designed to address these challenges. HYDRO utilizes an online simulator to complement the limited amount of offline datasets in the non-trivial context of robust RL. By measuring and minimizing performance gaps between the simulator and the worst-case models in the uncertainty set, HYDRO employs novel uncertainty filtering and prioritized sampling to select the most relevant and reliable simulator samples. Our extensive experiments demonstrate HYDRO's superior performance over existing methods across various tasks, underscoring its potential to improve sample efficiency in offline robust RL.
Authors: Kyle R. Chickering, Bangzheng Li, Muhao Chen
Abstract: Multimodal Large Language Models (MLLMs) encode images into visual tokens, aligning visual and textual signals within a shared latent space to facilitate crossmodal representation learning. The CLIP model is a widely adopted foundational vision language model whose vision encoder has played a critical role in the development of MLLMs such as LLaVA. However, the CLIP vision encoder suffers from notable limitations including being constrained to only handling fixed input resolutions and a failure to produce separated embeddings for dissimilar images. Replacing the vision encoder of an existing model typically incurs substantial computational costs because such a change often necessitates retraining the entire model pipeline. In this work, we identify two factors which underlie the limitations of the CLIP vision encoder: mesoscopic bias and interpolation bias. To address these issues, we propose QLIP, a drop-in replacement for CLIP that can be seamlessly integrated with existing MLLMs with only a few lines of code and can enhance both coarse-grained and fine-grained visual understanding, without re-training. QLIP is designed around an image quadtree which replaces the standard uniform grid patches with a novel content aware patchification. Our experimental results demonstrate that QLIP improves the general visual question answering accuracy of the LLaVA v1.5 model series across various model sizes--without requiring retraining or fine-tuning of the full MLLM. Notably, QLIP boosts detailed understanding performance on the challenging $V^{\ast}$ benchmark by up to 13.6 percent.
Authors: Ruskin Raj Manku, Yuzhi Tang, Xingjian Shi, Mu Li, Alex Smola
Abstract: Text-to-Speech (TTS) benchmarks often fail to capture how well models handle nuanced and semantically complex text. Building on $\textit{EmergentTTS}$, we introduce $\textit{EmergentTTS-Eval}$, a comprehensive benchmark covering six challenging TTS scenarios: emotions, paralinguistics, foreign words, syntactic complexity, complex pronunciation (e.g. URLs, formulas), and questions. Crucially, our framework automates both test-case generation and evaluation, making the benchmark easily extensible. Starting from a small set of human-written seed prompts, we iteratively extend them using LLMs to target specific structural, phonetic and prosodic challenges, resulting in 1,645 diverse test cases. Moreover, we employ a model-as-a-judge approach, using a Large Audio Language Model (LALM) to assess the speech across multiple dimensions such as expressed emotion, prosodic, intonational, and pronunciation accuracy. We evaluate state-of-the-art open-source and proprietary TTS systems, such as 11Labs, Deepgram, and OpenAI's 4o-mini-TTS, on EmergentTTS-Eval, demonstrating its ability to reveal fine-grained performance differences. Results show that the model-as-a-judge approach offers robust TTS assessment and a high correlation with human preferences. We open source the evaluation $\href{https://github.com/boson-ai/EmergentTTS-Eval-public}{code}$ and the $\href{https://huggingface.co/datasets/bosonai/EmergentTTS-Eval}{dataset}$.
URLs: https://github.com/boson-ai/EmergentTTS-Eval-public, https://huggingface.co/datasets/bosonai/EmergentTTS-Eval
Authors: Liangkai Hang, Junjie Yao, Zhiwei Bai, Tianyi Chen, Yang Chen, Rongjie Diao, Hezhou Li, Pengxiao Lin, Zhiwei Wang, Cheng Xu, Zhongwang Zhang, Zhangchen Zhou, Zhiyu Li, Zehao Lin, Kai Chen, Feiyu Xiong, Yaoyu Zhang, Weinan E, Hongkang Yang, Zhi-Qin John Xu
Abstract: The reasoning ability of large language models (LLMs) has been rapidly advancing in recent years, attracting interest in more fundamental approaches that can reliably enhance their generalizability. This work demonstrates that model complexity control, conveniently implementable by adjusting the initialization rate and weight decay coefficient, improves the scaling law of LLMs consistently over varying model sizes and data sizes. This gain is further illustrated by comparing the benchmark performance of 2.4B models pretrained on 1T tokens with different complexity hyperparameters. Instead of fixing the initialization std, we found that a constant initialization rate (the exponent of std) enables the scaling law to descend faster in both model and data sizes. These results indicate that complexity control is a promising direction for the continual advancement of LLMs.
Authors: Juwei Yue, Haikuo Li, Jiawei Sheng, Xiaodong Li, Taoyu Su, Tingwen Liu, Li Guo
Abstract: Graph neural networks (GNNs) leverage message passing mechanisms to learn the topological features of graph data. Traditional GNNs learns node features in a spatial domain unrelated to the topology, which can hardly ensure topological features. In this paper, we formulates message passing as a system of hyperbolic partial differential equations (hyperbolic PDEs), constituting a dynamical system that explicitly maps node representations into a particular solution space. This solution space is spanned by a set of eigenvectors describing the topological structure of graphs. Within this system, for any moment in time, a node features can be decomposed into a superposition of the basis of eigenvectors. This not only enhances the interpretability of message passing but also enables the explicit extraction of fundamental characteristics about the topological structure. Furthermore, by solving this system of hyperbolic partial differential equations, we establish a connection with spectral graph neural networks (spectral GNNs), serving as a message passing enhancement paradigm for spectral GNNs.We further introduce polynomials to approximate arbitrary filter functions. Extensive experiments demonstrate that the paradigm of hyperbolic PDEs not only exhibits strong flexibility but also significantly enhances the performance of various spectral GNNs across diverse graph tasks.
Authors: Xingjian Wu, Xiangfei Qiu, Hongfan Gao, Jilin Hu, Bin Yang, Chenjuan Guo
Abstract: Probabilistic Time Series Forecasting (PTSF) plays a crucial role in decision-making across various fields, including economics, energy, and transportation. Most existing methods excell at short-term forecasting, while overlooking the hurdles of Long-term Probabilistic Time Series Forecasting (LPTSF). As the forecast horizon extends, the inherent nonlinear dynamics have a significant adverse effect on prediction accuracy, and make generative models inefficient by increasing the cost of each iteration. To overcome these limitations, we introduce $K^2$VAE, an efficient VAE-based generative model that leverages a KoopmanNet to transform nonlinear time series into a linear dynamical system, and devises a KalmanNet to refine predictions and model uncertainty in such linear system, which reduces error accumulation in long-term forecasting. Extensive experiments demonstrate that $K^2$VAE outperforms state-of-the-art methods in both short- and long-term PTSF, providing a more efficient and accurate solution.
Authors: Yinghao Tang, Tingfeng Lan, Xiuqi Huang, Hui Lu, Wei Chen
Abstract: Existing Large Language Model (LLM) serving systems prioritize maximum throughput. They often neglect Service Level Objectives (SLOs) such as Time to First Token (TTFT) and Time Per Output Token (TPOT), which leads to suboptimal SLO attainment. This paper introduces SCORPIO, an SLO-oriented LLM serving system designed to maximize system goodput and SLO attainment for workloads with heterogeneous SLOs. Our core insight is to exploit SLO heterogeneity for adaptive scheduling across admission control, queue management, and batch selection. SCORPIO features a TTFT Guard, which employs least-deadline-first reordering and rejects unattainable requests, and a TPOT Guard, which utilizes a VBS-based admission control and a novel credit-based batching mechanism. Both guards are supported by a predictive module. Evaluations demonstrate that SCORPIO improves system goodput by up to 14.4X and SLO adherence by up to 46.5% compared to state-of-the-art baselines.
Authors: Zhihao Wang, Wenke Huang, Tian Chen, Zekun Shi, Guancheng Wan, Yu Qiao, Bin Yang, Jian Wang, Bing Li, Mang Ye
Abstract: The Vision Language Model (VLM) excels in aligning vision and language representations, and prompt learning has emerged as a key technique for adapting such models to downstream tasks. However, the application of prompt learning with VLM in federated learning (\fl{}) scenarios remains underexplored. This paper systematically investigates the behavioral differences between language prompt learning (LPT) and vision prompt learning (VPT) under data heterogeneity challenges, including label skew and domain shift. We conduct extensive experiments to evaluate the impact of various \fl{} and prompt configurations, such as client scale, aggregation strategies, and prompt length, to assess the robustness of Federated Prompt Learning (FPL). Furthermore, we explore strategies for enhancing prompt learning in complex scenarios where label skew and domain shift coexist, including leveraging both prompt types when computational resources allow. Our findings offer practical insights into optimizing prompt learning in federated settings, contributing to the broader deployment of VLMs in privacy-preserving environments.
Authors: Minh Nguyen Nhat To, Paul F RWilson, Viet Nguyen, Mohamed Harmanani, Michael Cooper, Fahimeh Fooladgar, Purang Abolmaesumi, Parvin Mousavi, Rahul G. Krishnan
Abstract: The subpopulationtion shift, characterized by a disparity in subpopulation distributibetween theween the training and target datasets, can significantly degrade the performance of machine learning models. Current solutions to subpopulation shift involve modifying empirical risk minimization with re-weighting strategies to improve generalization. This strategy relies on assumptions about the number and nature of subpopulations and annotations on group membership, which are unavailable for many real-world datasets. Instead, we propose using an ensemble of diverse classifiers to adaptively capture risk associated with subpopulations. Given a feature extractor network, we replace its standard linear classification layer with a mixture of prototypical classifiers, where each member is trained to classify the data while focusing on different features and samples from other members. In empirical evaluation on nine real-world datasets, covering diverse domains and kinds of subpopulation shift, our method of Diverse Prototypical Ensembles (DPEs) often outperforms the prior state-of-the-art in worst-group accuracy. The code is available at https://github.com/minhto2802/dpe4subpop
Authors: Dongwoo Lee, Dong Bok Lee, Steven Adriaensen, Juho Lee, Sung Ju Hwang, Frank Hutter, Seon Joo Kim, Hae Beom Lee
Abstract: Scaling has been a major driver of recent advancements in deep learning. Numerous empirical studies have found that scaling laws often follow the power-law and proposed several variants of power-law functions to predict the scaling behavior at larger scales. However, existing methods mostly rely on point estimation and do not quantify uncertainty, which is crucial for real-world applications involving decision-making problems such as determining the expected performance improvements achievable by investing additional computational resources. In this work, we explore a Bayesian framework based on Prior-data Fitted Networks (PFNs) for neural scaling law extrapolation. Specifically, we design a prior distribution that enables the sampling of infinitely many synthetic functions resembling real-world neural scaling laws, allowing our PFN to meta-learn the extrapolation. We validate the effectiveness of our approach on real-world neural scaling laws, comparing it against both the existing point estimation methods and Bayesian approaches. Our method demonstrates superior performance, particularly in data-limited scenarios such as Bayesian active learning, underscoring its potential for reliable, uncertainty-aware extrapolation in practical applications.
Authors: Siwen Wang, Shitou Zhang, Wan-Lin Chen, Dung Truong, Tzyy-Ping Jung
Abstract: Recent advancements in Large Language Models have inspired the development of foundation models across various domains. In this study, we evaluate the efficacy of Large EEG Models (LEMs) by fine-tuning LaBraM, a state-of-the-art foundation EEG model, on a real-world stress classification dataset collected in a graduate classroom. Unlike previous studies that primarily evaluate LEMs using data from controlled clinical settings, our work assesses their applicability to real-world environments. We train a binary classifier that distinguishes between normal and elevated stress states using resting-state EEG data recorded from 18 graduate students during a class session. The best-performing fine-tuned model achieves a balanced accuracy of 90.47% with a 5-second window, significantly outperforming traditional stress classifiers in both accuracy and inference efficiency. We further evaluate the robustness of the fine-tuned LEM under random data shuffling and reduced channel counts. These results demonstrate the capability of LEMs to effectively process real-world EEG data and highlight their potential to revolutionize brain-computer interface applications by shifting the focus from model-centric to data-centric design.
Authors: Tianci Bu, Le Zhou, Wenchuan Yang, Jianhong Mou, Kang Yang, Suoyi Tan, Feng Yao, Jingyuan Wang, Xin Lu
Abstract: Trajectory data is crucial for various applications but often suffers from incompleteness due to device limitations and diverse collection scenarios. Existing imputation methods rely on sparse trajectory or travel information, such as velocity, to infer missing points. However, these approaches assume that sparse trajectories retain essential behavioral patterns, which place significant demands on data acquisition and overlook the potential of large-scale human trajectory embeddings. To address this, we propose ProDiff, a trajectory imputation framework that uses only two endpoints as minimal information. It integrates prototype learning to embed human movement patterns and a denoising diffusion probabilistic model for robust spatiotemporal reconstruction. Joint training with a tailored loss function ensures effective imputation. ProDiff outperforms state-of-the-art methods, improving accuracy by 6.28\% on FourSquare and 2.52\% on WuXi. Further analysis shows a 0.927 correlation between generated and real trajectories, demonstrating the effectiveness of our approach.
Authors: Tianteng Gu, Bei Liu, Bo Xiao, Ke Zeng, Jiacheng Liu, Yanmin Qian
Abstract: Pruning is a widely used technique to compress large language models (LLMs) by removing unimportant weights, but it often suffers from significant performance degradation - especially under semi-structured sparsity constraints. Existing pruning methods primarily focus on estimating the importance of individual weights, which limits their ability to preserve critical capabilities of the model. In this work, we propose a new perspective: rather than merely selecting which weights to prune, we first redistribute parameter importance to make the model inherently more amenable to pruning. By minimizing the information entropy of normalized importance scores, our approach concentrates importance onto a smaller subset of weights, thereby enhancing pruning robustness. We instantiate this idea through DenoiseRotator, which applies learnable orthogonal transformations to the model's weight matrices. Our method is model-agnostic and can be seamlessly integrated with existing pruning techniques such as Magnitude, SparseGPT, and Wanda. Evaluated on LLaMA3, Qwen2.5, and Mistral models under 50% unstructured and 2:4 semi-structured sparsity, DenoiseRotator consistently improves perplexity and zero-shot accuracy. For instance, on LLaMA3-70B pruned with SparseGPT at 2:4 semi-structured sparsity, DenoiseRotator reduces the perplexity gap to the dense model by 58%, narrowing the degradation from 8.1 to 3.4 points. Codes are available at https://github.com/Axel-gu/DenoiseRotator.
Authors: Zhen Xiang, Aliyah R. Hsu, Austin V. Zane, Aaron E. Kornblith, Margaret J. Lin-Martore, Jasmanpreet C. Kaur, Vasuda M. Dokiparthi, Bo Li, Bin Yu
Abstract: Clinical decision-making is inherently complex and fast-paced, particularly in emergency departments (EDs) where critical, rapid and high-stakes decisions are made. Clinical Decision Rules (CDRs) are standardized evidence-based tools that combine signs, symptoms, and clinical variables into decision trees to make consistent and accurate diagnoses. CDR usage is often hindered by the clinician's cognitive load, limiting their ability to quickly recall and apply the appropriate rules. We introduce CDR-Agent, a novel LLM-based system designed to enhance ED decision-making by autonomously identifying and applying the most appropriate CDRs based on unstructured clinical notes. To validate CDR-Agent, we curated two novel ED datasets: synthetic and CDR-Bench, although CDR-Agent is applicable to non ED clinics. CDR-Agent achieves a 56.3\% (synthetic) and 8.7\% (CDR-Bench) accuracy gain relative to the standalone LLM baseline in CDR selection. Moreover, CDR-Agent significantly reduces computational overhead. Using these datasets, we demonstrated that CDR-Agent not only selects relevant CDRs efficiently, but makes cautious yet effective imaging decisions by minimizing unnecessary interventions while successfully identifying most positively diagnosed cases, outperforming traditional LLM prompting approaches. Code for our work can be found at: https://github.com/zhenxianglance/medagent-cdr-agent
Authors: Tarun Suresh, Debangshu Banerjee, Shubham Ugare, Sasa Misailovic, Gagandeep Singh
Abstract: Diffusion LLMs have emerged as a promising alternative to conventional autoregressive LLMs, offering significant potential for improved runtime efficiency. However, existing diffusion models lack the ability to provably enforce user-specified formal constraints, such as regular expressions, which makes them unreliable for tasks that require structured outputs, such as fixed-schema JSON generation. Unlike autoregressive models that generate tokens sequentially, diffusion LLMs predict a block of tokens in parallel. This parallelism makes traditional constrained decoding algorithms, which are designed for sequential token prediction, ineffective at preserving the true output distribution. To address this limitation, we propose DINGO, a dynamic programming-based constrained decoding strategy that is both efficient and provably distribution-preserving. DINGO enables sampling of output strings with the highest probability under the model's predicted distribution, while strictly satisfying any user-specified regular expression. On standard symbolic math and JSON generation benchmarks, DINGO achieves up to a 68 percentage point improvement over unconstrained inference
Authors: Lingkai Kong, Haichuan Wang, Tonghan Wang, Guojun Xiong, Milind Tambe
Abstract: Incorporating pre-collected offline data from a source environment can significantly improve the sample efficiency of reinforcement learning (RL), but this benefit is often challenged by discrepancies between the transition dynamics of the source and target environments. Existing methods typically address this issue by penalizing or filtering out source transitions in high dynamics-gap regions. However, their estimation of the dynamics gap often relies on KL divergence or mutual information, which can be ill-defined when the source and target dynamics have disjoint support. To overcome these limitations, we propose CompFlow, a method grounded in the theoretical connection between flow matching and optimal transport. Specifically, we model the target dynamics as a conditional flow built upon the output distribution of the source-domain flow, rather than learning it directly from a Gaussian prior. This composite structure offers two key advantages: (1) improved generalization for learning target dynamics, and (2) a principled estimation of the dynamics gap via the Wasserstein distance between source and target transitions. Leveraging our principled estimation of the dynamics gap, we further introduce an optimistic active data collection strategy that prioritizes exploration in regions of high dynamics gap, and theoretically prove that it reduces the performance disparity with the optimal policy. Empirically, CompFlow outperforms strong baselines across several RL benchmarks with shifted dynamics.
Authors: Denis Mamba Kabala, Adel Hafiane, Laurent Bobelin, Raphael Canals
Abstract: Crop disease detection and classification is a critical challenge in agriculture, with major implications for productivity, food security, and environmental sustainability. While deep learning models such as CNN and ViT have shown excellent performance in classifying plant diseases from images, their large-scale deployment is often limited by data privacy concerns. Federated Learning (FL) addresses this issue, but centralized FL remains vulnerable to single-point failures and scalability limits. In this paper, we introduce a novel Decentralized Federated Learning (DFL) framework that uses validation loss (Loss_val) both to guide model sharing between peers and to correct local training via an adaptive loss function controlled by weighting parameter. We conduct extensive experiments using PlantVillage datasets with three deep learning architectures (ResNet50, VGG16, and ViT_B16), analyzing the impact of weighting parameter, the number of shared models, the number of clients, and the use of Loss_val versus Loss_train of other clients. Results demonstrate that our DFL approach not only improves accuracy and convergence speed, but also ensures better generalization and robustness across heterogeneous data environments making it particularly well-suited for privacy-preserving agricultural applications.
Authors: Peizheng Guo, Jingyao Wang, Huijie Guo, Jiangmeng Li, Chuxiong Sun, Changwen Zheng, Wenwen Qiang
Abstract: Multi-Modal Learning (MML) integrates information from diverse modalities to improve predictive accuracy. However, existing methods mainly aggregate gradients with fixed weights and treat all dimensions equally, overlooking the intrinsic gradient uncertainty of each modality. This may lead to (i) excessive updates in sensitive dimensions, degrading performance, and (ii) insufficient updates in less sensitive dimensions, hindering learning. To address this issue, we propose BOGC-MML, a Bayesian-Oriented Gradient Calibration method for MML to explicitly model the gradient uncertainty and guide the model optimization towards the optimal direction. Specifically, we first model each modality's gradient as a random variable and derive its probability distribution, capturing the full uncertainty in the gradient space. Then, we propose an effective method that converts the precision (inverse variance) of each gradient distribution into a scalar evidence. This evidence quantifies the confidence of each modality in every gradient dimension. Using these evidences, we explicitly quantify per-dimension uncertainties and fuse them via a reduced Dempster-Shafer rule. The resulting uncertainty-weighted aggregation produces a calibrated update direction that balances sensitivity and conservatism across dimensions. Extensive experiments on multiple benchmark datasets demonstrate the effectiveness and advantages of the proposed method.
Authors: Chang Yu, Fang Liu, Jie Zhu, Shaobo Guo, Yifan Gao, Zhongheng Yang, Meiwei Liu, Qianwen Xing
Abstract: This paper proposes a hybrid framework combining LSTM (Long Short-Term Memory) networks with LightGBM and CatBoost for stock price prediction. The framework processes time-series financial data and evaluates performance using seven models: Artificial Neural Networks (ANNs), Convolutional Neural Networks (CNNs), Bidirectional LSTM (BiLSTM), vanilla LSTM, XGBoost, LightGBM, and standard Neural Networks (NNs). Key metrics, including MAE, R-squared, MSE, and RMSE, are used to establish benchmarks across different time scales. Building on these benchmarks, we develop an ensemble model that combines the strengths of sequential and tree-based approaches. Experimental results show that the proposed framework improves accuracy by 10 to 15 percent compared to individual models and reduces error during market changes. This study highlights the potential of ensemble methods for financial forecasting and provides a flexible design for integrating new machine learning techniques.
Authors: Junyi An, Xinyu Lu, Chao Qu, Yunfei Shi, Peijia Lin, Qianwei Tang, Licheng Xu, Fenglei Cao, Yuan Qi
Abstract: SE(3)-equivariant Graph Neural Networks (GNNs) have significantly advanced molecular system modeling by employing group representations. However, their message passing processes, which rely on tensor product-based convolutions, are limited by insufficient non-linearity and incomplete group representations, thereby restricting expressiveness. To overcome these limitations, we introduce the Equivariant Spherical Transformer (EST), a novel framework that leverages a Transformer structure within the spatial domain of group representations after Fourier transform. We theoretically and empirically demonstrate that EST can encompass the function space of tensor products while achieving superior expressiveness. Furthermore, EST's equivariant inductive bias is guaranteed through a uniform sampling strategy for the Fourier transform. Our experiments demonstrate state-of-the-art performance by EST on various molecular benchmarks, including OC20 and QM9.
Authors: Chongjie Si, Zhiyi Shi, Yadao Wang, Xiaokang Yang, Susanto Rahardja, Wei Shen
Abstract: The rapid development of large language models has revolutionized natural language processing, but their fine-tuning remains computationally expensive, hindering broad deployment. Parameter-efficient fine-tuning (PEFT) methods, such as LoRA, have emerged as solutions. Recent work like DoRA attempts to further decompose weight adaptation into direction and magnitude components. However, existing formulations often define direction heuristically at the column level, lacking a principled geometric foundation. In this paper, we propose MAP, a novel framework that reformulates weight matrices as high-dimensional vectors and decouples their adaptation into direction and magnitude in a rigorous manner. MAP normalizes the pre-trained weights, learns a directional update, and introduces two scalar coefficients to independently scale the magnitude of the base and update vectors. This design enables more interpretable and flexible adaptation, and can be seamlessly integrated into existing PEFT methods. Extensive experiments show that MAP significantly improves performance when coupling with existing methods, offering a simple yet powerful enhancement to existing PEFT methods. Given the universality and simplicity of MAP, we hope it can serve as a default setting for designing future PEFT methods.
Authors: Kuan Xu, Zhiguang Cao, Chenlong Zheng, Linong Liu
Abstract: In this study, we propose a reinforcement learning-based adaptive variable neighborhood search (RL-AVNS) method designed for effectively solving the Vehicle Routing Problem with Multiple Time Windows (VRPMTW). Unlike traditional adaptive approaches that rely solely on historical operator performance, our method integrates a reinforcement learning framework to dynamically select neighborhood operators based on real-time solution states and learned experience. We introduce a fitness metric that quantifies customers' temporal flexibility to improve the shaking phase, and employ a transformer-based neural policy network to intelligently guide operator selection during the local search. Extensive computational experiments are conducted on realistic scenarios derived from the replenishment of unmanned vending machines, characterized by multiple clustered replenishment windows. Results demonstrate that RL-AVNS significantly outperforms traditional variable neighborhood search (VNS), adaptive VNS (AVNS), and state-of-the-art learning-based heuristics, achieving substantial improvements in solution quality and computational efficiency across various instance scales and time window complexities. Particularly notable is the algorithm's capability to generalize effectively to problem instances not encountered during training, underscoring its practical utility for complex logistics scenarios.
Authors: Chongjie Si, Xuankun Yang, Muqing Liu, Yadao Wang, Xiaokang Yang, Wenbo Su, Bo Zheng, Wei Shen
Abstract: Large-scale foundation models have demonstrated remarkable versatility across a wide range of downstream tasks. However, fully fine-tuning these models incurs prohibitive computational costs, motivating the development of Parameter-Efficient Fine-Tuning (PEFT) methods such as LoRA, which introduces low-rank updates to pre-trained weights. Despite their empirical success, the underlying mechanisms by which PEFT modifies model parameters remain underexplored. In this work, we present a systematic investigation into the structural changes of weight matrices during fully fine-tuning. Through singular value decomposition (SVD), we reveal that fine-tuning predominantly amplifies the top singular values while leaving the remainder largely intact, suggesting that task-specific knowledge is injected into a low-dimensional subspace. Furthermore, we find that the dominant singular vectors are reoriented in task-specific directions, whereas the non-dominant subspace remains stable. Building on these insights, we propose a novel method that leverages learnable rescaling of top singular directions, enabling precise modulation of the most influential components without disrupting the global structure. Our approach achieves consistent improvements over strong baselines across multiple tasks, highlighting the efficacy of structurally informed fine-tuning.
Authors: Abhishek Vijaya Kumar, Eric Ding, Arjun Devraj, Darius Bunandar, Rachee Singh
Abstract: When accelerators fail in modern ML datacenters, operators migrate the affected ML training or inference jobs to entirely new racks. This approach, while preserving network performance, is highly inefficient, requiring datacenters to reserve full racks of idle accelerators for fault tolerance. In this paper, we address this resource inefficiency by introducing LUMION, a novel reconfigurable optical fabric for connecting accelerators within a datacenter rack. Instead of migrating entire ML jobs, LUMION dynamically integrates spare accelerators into ongoing workloads as failures occur, thereby maintaining consistent performance without costly migrations. We show the benefits of LUMION by building an end-to-end hardware prototype. Our experiments fine-tune Llama 3.2 and show that LUMION swaps a failed GPU with a healthy one and restarts the ML job within ~ 1 second of the failure. LUMION achieves higher inter-GPU bandwidth compared to traditional electrical racks after replacing failed accelerators with spare ones, leading to nearly 2X improvement in fine-tuning throughput.
Authors: Ning Liu, Yue Yu
Abstract: Attention mechanisms have emerged as transformative tools in core AI domains such as natural language processing and computer vision. Yet, their largely untapped potential for modeling intricate physical systems presents a compelling frontier. Learning such systems often entails discovering operators that map between functional spaces using limited instances of function pairs -- a task commonly framed as a severely ill-posed inverse PDE problem. In this work, we introduce Neural Interpretable PDEs (NIPS), a novel neural operator architecture that builds upon and enhances Nonlocal Attention Operators (NAO) in both predictive accuracy and computational efficiency. NIPS employs a linear attention mechanism to enable scalable learning and integrates a learnable kernel network that acts as a channel-independent convolution in Fourier space. As a consequence, NIPS eliminates the need to explicitly compute and store large pairwise interactions, effectively amortizing the cost of handling spatial interactions into the Fourier transform. Empirical evaluations demonstrate that NIPS consistently surpasses NAO and other baselines across diverse benchmarks, heralding a substantial leap in scalable, interpretable, and efficient physics learning. Our code and data accompanying this paper are available at https://github.com/fishmoon1234/Nonlocal-Attention-Operator.
URLs: https://github.com/fishmoon1234/Nonlocal-Attention-Operator.
Authors: Pengfei Zhou, Yunlong Liu, Junli Liang, Qi Song, Xiangyang Li
Abstract: Time series forecasting with exogenous variables is a critical emerging paradigm that presents unique challenges in modeling dependencies between variables. Traditional models often struggle to differentiate between endogenous and exogenous variables, leading to inefficiencies and overfitting. In this paper, we introduce CrossLinear, a novel Linear-based forecasting model that addresses these challenges by incorporating a plug-and-play cross-correlation embedding module. This lightweight module captures the dependencies between variables with minimal computational cost and seamlessly integrates into existing neural networks. Specifically, it captures time-invariant and direct variable dependencies while disregarding time-varying or indirect dependencies, thereby mitigating the risk of overfitting in dependency modeling and contributing to consistent performance improvements. Furthermore, CrossLinear employs patch-wise processing and a global linear head to effectively capture both short-term and long-term temporal dependencies, further improving its forecasting precision. Extensive experiments on 12 real-world datasets demonstrate that CrossLinear achieves superior performance in both short-term and long-term forecasting tasks. The ablation study underscores the effectiveness of the cross-correlation embedding module. Additionally, the generalizability of this module makes it a valuable plug-in for various forecasting tasks across different domains. Codes are available at https://github.com/mumiao2000/CrossLinear.
Authors: Yuatyong Chaichana, Thanapat Trachu, Peerat Limkonchotiwat, Konpat Preechakul, Tirasan Khandhawit, Ekapol Chuangsuwanich
Abstract: In the era of large-scale training, model merging has evolved into a tool for creating multitasking models efficiently. It enables the knowledge of models to be fused, without the need for heavy computation as required in traditional multitask learning. Existing merging methods often assume that entries at identical positions in weight matrices serve the same function, enabling straightforward entry-wise comparison and merging. However, this assumption overlooks the complexity of finetuned neural networks, where neurons may develop distinct feature compositions, making direct entry-wise merging problematic. We present Decom-Renorm-Merge (DRM), a simple yet effective approach that leverages Singular Value Decomposition to decompose and coordinate weight matrices into an aligned joint space, where entry-wise merging becomes possible. We showcase the effectiveness of DRM across various settings ranging from smaller encoder-based such as ViT and DeBERTa, encoder-decoder-based such as T5, and larger decoder-based such as Llama3.1-8B. Our experimental results show that DRM outperforms several state-of-the-art merging techniques across full finetuning and low-rank adaptation settings. Moreover, our analysis reveals renormalization as the crucial component for creating a robust and even joint space for merging, significantly contributing to the method's performance.
Authors: Xinyu Yao, Daniel Bourgeois, Abhinav Jain, Yuxin Tang, Jiawen Yao, Zhimin Ding, Arlei Silva, Chris Jermaine
Abstract: We study the problem of assigning operations in a dataflow graph to devices to minimize execution time in a work-conserving system, with emphasis on complex machine learning workloads. Prior learning-based methods often struggle due to three key limitations: (1) reliance on bulk-synchronous systems like TensorFlow, which under-utilize devices due to barrier synchronization; (2) lack of awareness of the scheduling mechanism of underlying systems when designing learning-based methods; and (3) exclusive dependence on reinforcement learning, ignoring the structure of effective heuristics designed by experts. In this paper, we propose \textsc{Doppler}, a three-stage framework for training dual-policy networks consisting of 1) a $\mathsf{SEL}$ policy for selecting operations and 2) a $\mathsf{PLC}$ policy for placing chosen operations on devices. Our experiments show that \textsc{Doppler} outperforms all baseline methods across tasks by reducing system execution time and additionally demonstrates sampling efficiency by reducing per-episode training time.
Authors: Zhe Ye, Zhengxu Yan, Jingxuan He, Timothe Kasriel, Kaiyu Yang, Dawn Song
Abstract: Large language models (LLMs) are increasingly integrated in software development, but ensuring correctness in LLM-generated code remains challenging and often requires costly manual review. Verifiable code generation -- jointly generating code, specifications, and proofs of code-specification alignment -- offers a promising path to address this limitation and further unleash LLMs' benefits in coding. Yet, there exists a significant gap in evaluation: current benchmarks often lack support for end-to-end verifiable code generation. In this paper, we introduce Verina (Verifiable Code Generation Arena), a high-quality benchmark enabling a comprehensive and modular evaluation of code, specification, and proof generation as well as their compositions. Verina consists of 189 manually curated coding tasks in Lean, with detailed problem descriptions, reference implementations, formal specifications, and extensive test suites. Our extensive evaluation of state-of-the-art LLMs reveals significant challenges in verifiable code generation, especially in proof generation, underscoring the need for improving LLM-based theorem provers in verification domains. The best model, OpenAI o4-mini, generates only 61.4% correct code, 51.0% sound and complete specifications, and 3.6% successful proofs, with one trial per task. We hope Verina will catalyze progress in verifiable code generation by providing a rigorous and comprehensive benchmark. We release our dataset on https://huggingface.co/datasets/sunblaze-ucb/verina and our evaluation code on https://github.com/sunblaze-ucb/verina.
URLs: https://huggingface.co/datasets/sunblaze-ucb/verina, https://github.com/sunblaze-ucb/verina.
Authors: Michal Nauman, Marek Cygan, Carmelo Sferrazza, Aviral Kumar, Pieter Abbeel
Abstract: Recent advances in language modeling and vision stem from training large models on diverse, multi-task data. This paradigm has had limited impact in value-based reinforcement learning (RL), where improvements are often driven by small models trained in a single-task context. This is because in multi-task RL sparse rewards and gradient conflicts make optimization of temporal difference brittle. Practical workflows for generalist policies therefore avoid online training, instead cloning expert trajectories or distilling collections of single-task policies into one agent. In this work, we show that the use of high-capacity value models trained via cross-entropy and conditioned on learnable task embeddings addresses the problem of task interference in online RL, allowing for robust and scalable multi-task training. We test our approach on 7 multi-task benchmarks with over 280 unique tasks, spanning high degree-of-freedom humanoid control and discrete vision-based RL. We find that, despite its simplicity, the proposed approach leads to state-of-the-art single and multi-task performance, as well as sample-efficient transfer to new tasks.
Authors: Le Yang, Vincent Y. F. Tan, Wang Chi Cheung
Abstract: We study the best arm identification (BAI) problem with potentially biased offline data in the fixed confidence setting, which commonly arises in real-world scenarios such as clinical trials. We prove an impossibility result for adaptive algorithms without prior knowledge of the bias bound between online and offline distributions. To address this, we propose the LUCB-H algorithm, which introduces adaptive confidence bounds by incorporating an auxiliary bias correction to balance offline and online data within the LUCB framework. Theoretical analysis shows that LUCB-H matches the sample complexity of standard LUCB when offline data is misleading and significantly outperforms it when offline data is helpful. We also derive an instance-dependent lower bound that matches the upper bound of LUCB-H in certain scenarios. Numerical experiments further demonstrate the robustness and adaptability of LUCB-H in effectively incorporating offline data.
Authors: Shohei Enomoto
Abstract: Deep learning models often struggle to maintain performance when deployed on data distributions different from their training data, particularly in real-world applications where environmental conditions frequently change. While Multi-source Domain Generalization (MDG) has shown promise in addressing this challenge by leveraging multiple source domains during training, its practical application is limited by the significant costs and difficulties associated with creating multi-domain datasets. To address this limitation, we propose Pseudo Multi-source Domain Generalization (PMDG), a novel framework that enables the application of sophisticated MDG algorithms in more practical Single-source Domain Generalization (SDG) settings. PMDG generates multiple pseudo-domains from a single source domain through style transfer and data augmentation techniques, creating a synthetic multi-domain dataset that can be used with existing MDG algorithms. Through extensive experiments with PseudoDomainBed, our modified version of the DomainBed benchmark, we analyze the effectiveness of PMDG across multiple datasets and architectures. Our analysis reveals several key findings, including a positive correlation between MDG and PMDG performance and the potential of pseudo-domains to match or exceed actual multi-domain performance with sufficient data. These comprehensive empirical results provide valuable insights for future research in domain generalization. Our code is available at https://github.com/s-enmt/PseudoDomainBed.
Authors: Shiwei Li, Xiandi Luo, Haozhao Wang, Xing Tang, Shijie Xu, Weihong Luo, Yuhua Li, Xiuqiang He, Ruixuan Li
Abstract: To improve the training efficiency of federated learning (FL), previous research has employed low-rank decomposition techniques to reduce communication overhead. In this paper, we seek to enhance the performance of these low-rank decomposition methods. Specifically, we focus on three key issues related to decomposition in FL: what to decompose, how to decompose, and how to aggregate. Subsequently, we introduce three novel techniques: Model Update Decomposition (MUD), Block-wise Kronecker Decomposition (BKD), and Aggregation-Aware Decomposition (AAD), each targeting a specific issue. These techniques are complementary and can be applied simultaneously to achieve optimal performance. Additionally, we provide a rigorous theoretical analysis to ensure the convergence of the proposed MUD. Extensive experimental results show that our approach achieves faster convergence and superior accuracy compared to relevant baseline methods. The code is available at https://github.com/Leopold1423/fedmud-icml25.
Authors: Tian Tian, Chunyan Miao, Hangwei Qian
Abstract: Contrastive learning has emerged as a competent approach for unsupervised representation learning. However, the design of an optimal augmentation strategy, although crucial for contrastive learning, is less explored for time series classification tasks. Existing predefined time-domain augmentation methods are primarily adopted from vision and are not specific to time series data. Consequently, this cross-modality incompatibility may distort the semantically relevant information of time series by introducing mismatched patterns into the data. To address this limitation, we present a novel perspective from the frequency domain and identify three advantages for downstream classification: global, independent, and compact. To fully utilize the three properties, we propose the lightweight yet effective Frequency Refined Augmentation (FreRA) tailored for time series contrastive learning on classification tasks, which can be seamlessly integrated with contrastive learning frameworks in a plug-and-play manner. Specifically, FreRA automatically separates critical and unimportant frequency components. Accordingly, we propose semantic-aware Identity Modification and semantic-agnostic Self-adaptive Modification to protect semantically relevant information in the critical frequency components and infuse variance into the unimportant ones respectively. Theoretically, we prove that FreRA generates semantic-preserving views. Empirically, we conduct extensive experiments on two benchmark datasets, including UCR and UEA archives, as well as five large-scale datasets on diverse applications. FreRA consistently outperforms ten leading baselines on time series classification, anomaly detection, and transfer learning tasks, demonstrating superior capabilities in contrastive representation learning and generalization in transfer learning scenarios across diverse datasets.
Authors: Srijith Nair, Michael Lin, Amirreza Talebi, Peizhong Ju, Elizabeth Bentley, Jia Liu
Abstract: Collaborative training methods like Federated Learning (FL) and Split Learning (SL) enable distributed machine learning without sharing raw data. However, FL assumes clients can train entire models, which is infeasible for large-scale models. In contrast, while SL alleviates the client memory constraint in FL by offloading most training to the server, it increases network latency due to its sequential nature. Other methods address the conundrum by using local loss functions for parallel client-side training to improve efficiency, but they lack server feedback and potentially suffer poor accuracy. We propose FSL-SAGE (Federated Split Learning via Smashed Activation Gradient Estimation), a new federated split learning algorithm that estimates server-side gradient feedback via auxiliary models. These auxiliary models periodically adapt to emulate server behavior on local datasets. We show that FSL-SAGE achieves a convergence rate of $\mathcal{O}(1/\sqrt{T})$, where $T$ is the number of communication rounds. This result matches FedAvg, while significantly reducing communication costs and client memory requirements. Our empirical results also verify that it outperforms existing state-of-the-art FSL methods, offering both communication efficiency and accuracy.
Authors: Hongcan Guo, Guoshun Nan, Yuan Yang, Diyang Zhang, Haotian Li, Zhican Chen, Qinchuan Zhou, Yuhan Ran, Xinye Cao, Sicong Leng, Xiaofeng Tao, Xudong Jiang
Abstract: Scaling Low-Rank Adaptation (LoRA)-based Mixture-of-Experts (MoE) facilitates large language models (LLMs) to efficiently adapt to diverse tasks. However, traditional gating mechanisms that route inputs to the best experts may fundamentally hinder LLMs' scalability, leading to poor generalization and underfitting issues. We identify that the root cause lies in the restricted expressiveness of existing weighted-sum mechanisms, both within and outside the convex cone of LoRA representations. This motivates us to propose RadarGate, a novel geometrically inspired gating method that introduces rotational operations of LoRAs representations to boost the expressiveness and facilitate richer feature interactions among multiple LoRAs for scalable LLMs. Specifically, we first fuse each LoRA representation to other LoRAs using a learnable component and then feed the output to a rotation matrix. This matrix involves learnable parameters that define the relative angular relationship between LoRA representations. Such a simple yet effective mechanism provides an extra degree of freedom, facilitating the learning of cross-LoRA synergies and properly tracking the challenging poor generalization and underfitting issues as the number of LoRA grows. Extensive experiments on 6 public benchmarks across 21 tasks show the effectiveness of our RadarGate for scaling LoRAs. We also provide valuable insights, revealing that the rotations to each pair of representations are contrastive, encouraging closer alignment of semantically similar representations during geometrical transformation while pushing distance ones further apart. We will release our code to the community.
Authors: Shahaf E. Finder, Ron Shapira Weber, Moshe Eliasof, Oren Freifeld, Eran Treister
Abstract: Message-Passing Neural Networks (MPNNs) have become a cornerstone for processing and analyzing graph-structured data. However, their effectiveness is often hindered by phenomena such as over-squashing, where long-range dependencies or interactions are inadequately captured and expressed in the MPNN output. This limitation mirrors the challenges of the Effective Receptive Field (ERF) in Convolutional Neural Networks (CNNs), where the theoretical receptive field is underutilized in practice. In this work, we show and theoretically explain the limited ERF problem in MPNNs. Furthermore, inspired by recent advances in ERF augmentation for CNNs, we propose an Interleaved Multiscale Message-Passing Neural Networks (IM-MPNN) architecture to address these problems in MPNNs. Our method incorporates a hierarchical coarsening of the graph, enabling message-passing across multiscale representations and facilitating long-range interactions without excessive depth or parameterization. Through extensive evaluations on benchmarks such as the Long-Range Graph Benchmark (LRGB), we demonstrate substantial improvements over baseline MPNNs in capturing long-range dependencies while maintaining computational efficiency.
Authors: Yekun Zhu, Min Tang, Zheng Ma
Abstract: In this study, we propose a novel neural network approach, termed DeepRTE, to address the steady-state Radiative Transfer Equation (RTE). The RTE is a differential-integral equation that governs the propagation of radiation through a participating medium, with applications spanning diverse domains such as neutron transport, atmospheric radiative transfer, heat transfer, and optical imaging. Our proposed DeepRTE framework leverages pre-trained attention-based neural networks to solve the RTE with high accuracy and computational efficiency. The efficacy of the proposed approach is substantiated through comprehensive numerical experiments.
Authors: Shiwei Li, Xiandi Luo, Xing Tang, Haozhao Wang, Hao Chen, Weihong Luo, Yuhua Li, Xiuqiang He, Ruixuan Li
Abstract: Low-rank adaptation (LoRA) is a widely used parameter-efficient fine-tuning method. In standard LoRA layers, one of the matrices, $A$ or $B$, is initialized to zero, ensuring that fine-tuning starts from the pretrained model. However, there is no theoretical support for this practice. In this paper, we investigate the impact of non-zero initialization on LoRA's fine-tuning dynamics from an infinite-width perspective. Our analysis reveals that, compared to zero initialization, simultaneously initializing $A$ and $B$ to non-zero values improves LoRA's robustness to suboptimal learning rates, particularly smaller ones. Further analysis indicates that although the non-zero initialization of $AB$ introduces random noise into the pretrained weight, it generally does not affect fine-tuning performance. In other words, fine-tuning does not need to strictly start from the pretrained model. The validity of our findings is confirmed through extensive experiments across various models and datasets. The code is available at https://github.com/Leopold1423/non_zero_lora-icml25.
Authors: Lifan Zhao, Yanyan Shen, Zhaoyang Liu, Xue Wang, Jiaji Deng
Abstract: Scaling laws motivate the development of Time Series Foundation Models (TSFMs) that pre-train vast parameters and achieve remarkable zero-shot forecasting performance. Surprisingly, even after fine-tuning, TSFMs cannot consistently outperform smaller, specialized models trained on full-shot downstream data. A key question is how to realize effective adaptation of TSFMs for a target forecasting task. Through empirical studies on various TSFMs, the pre-trained models often exhibit inherent sparsity and redundancy in computation, suggesting that TSFMs have learned to activate task-relevant network substructures to accommodate diverse forecasting tasks. To preserve this valuable prior knowledge, we propose a structured pruning method to regularize the subsequent fine-tuning process by focusing it on a more relevant and compact parameter space. Extensive experiments on seven TSFMs and six benchmarks demonstrate that fine-tuning a smaller, pruned TSFM significantly improves forecasting performance compared to fine-tuning original models. This "prune-then-finetune" paradigm often enables TSFMs to achieve state-of-the-art performance and surpass strong specialized baselines.
Authors: Xingyuan Pan, Chenlu Ye, Joseph Melkonian, Jiaqi W. Ma, Tong Zhang
Abstract: Training data attribution (TDA) methods aim to identify which training examples influence a model's predictions on specific test data most. By quantifying these influences, TDA supports critical applications such as data debugging, curation, and valuation. Gradient-based TDA methods rely on gradients and second-order information, limiting their applicability at scale. While recent random projection-based methods improve scalability, they often suffer from degraded attribution accuracy. Motivated by connections between uncertainty and influence functions, we introduce Daunce - a simple yet effective data attribution approach through uncertainty estimation. Our method operates by fine-tuning a collection of perturbed models and computing the covariance of per-example losses across these models as the attribution score. Daunce is scalable to large language models (LLMs) and achieves more accurate attribution compared to existing TDA methods. We validate Daunce on tasks ranging from vision tasks to LLM fine-tuning, and further demonstrate its compatibility with black-box model access. Applied to OpenAI's GPT models, our method achieves, to our knowledge, the first instance of data attribution on proprietary LLMs.
Authors: Fabiano Veglianti, Flavio Giorgi, Fabrizio Silvestri, Gabriele Tolomei
Abstract: In this work, we investigate the relationship between model generalization and counterfactual explainability in supervised learning. We introduce the notion of $\varepsilon$-valid counterfactual probability ($\varepsilon$-VCP) -- the probability of finding perturbations of a data point within its $\varepsilon$-neighborhood that result in a label change. We provide a theoretical analysis of $\varepsilon$-VCP in relation to the geometry of the model's decision boundary, showing that $\varepsilon$-VCP tends to increase with model overfitting. Our findings establish a rigorous connection between poor generalization and the ease of counterfactual generation, revealing an inherent trade-off between generalization and counterfactual explainability. Empirical results validate our theory, suggesting $\varepsilon$-VCP as a practical proxy for quantitatively characterizing overfitting.
Authors: Wanfu Gao, Jun Gao, Qingqi Han, Hanlin Pan, Kunpeng Liu
Abstract: The rapid growth in feature dimension may introduce implicit associations between features and labels in multi-label datasets, making the relationships between features and labels increasingly complex. Moreover, existing methods often adopt low-dimensional linear decomposition to explore the associations between features and labels. However, linear decomposition struggles to capture complex nonlinear associations and may lead to misalignment between the feature space and the label space. To address these two critical challenges, we propose innovative solutions. First, we design a random walk graph that integrates feature-feature, label-label, and feature-label relationships to accurately capture nonlinear and implicit indirect associations, while optimizing the latent representations of associations between features and labels after low-rank decomposition. Second, we align the variable spaces by leveraging low-dimensional representation coefficients, while preserving the manifold structure between the original high-dimensional multi-label data and the low-dimensional representation space. Extensive experiments and ablation studies conducted on seven benchmark datasets and three representative datasets using various evaluation metrics demonstrate the superiority of the proposed method\footnote{Code: https://github.com/Heilong623/-GRW-}.
Authors: Emo Todorov
Abstract: Policy gradients in continuous control have been derived for both stochastic and deterministic policies. Here we study the relationship between the two. In a widely-used family of MDPs involving Gaussian control noise and quadratic control costs, we show that the stochastic and deterministic policy gradients, natural gradients, and state value functions are identical; while the state-control value functions are different. We then develop a general procedure for constructing an MDP with deterministic policy that is equivalent to a given MDP with stochastic policy. The controls of this new MDP are the sufficient statistics of the stochastic policy in the original MDP. Our results suggest that policy gradient methods can be unified by approximating state value functions rather than state-control value functions.
Authors: Honoka Anada, Tatsuya Kaneko, Shinya Takamaeda-Yamazaki
Abstract: Federated learning (FL) enables multiple clients to collaboratively train models without sharing their data. Measuring participant contributions in FL is crucial for incentivizing clients and ensuring transparency. While various methods have been proposed for contribution measurement, they are designed exclusively for centralized federated learning (CFL), where a central server collects and aggregates client models, along with evaluating their contributions. Meanwhile, decentralized federated learning (DFL), in which clients exchange models directly without a central server, has gained significant attention for mitigating communication bottlenecks and eliminating a single point of failure. However, applying existing contribution measurement methods to DFL is challenging due to the presence of multiple global models and the absence of a central server. In this study, we present novel methodologies for measuring participant contributions in DFL. We first propose DFL-Shapley, an extension of the Shapley value tailored for DFL, adapting this widely used CFL metric to decentralized settings. Given the impracticality of computing the ideal DFL-Shapley in real-world systems, we introduce DFL-MR, a computable approximation that estimates overall contributions by accumulating round-wise Shapley values. We evaluate DFL-Shapley and DFL-MR across various FL scenarios and compare them with existing CFL metrics. The experimental results confirm DFL-Shapley as a valid ground-truth metric and demonstrate DFL-MR's proximity to DFL-Shapley across various settings, highlighting their effectiveness as contribution metrics in DFL.
Authors: Zonglin Yang, Zhexuan Gu, Houduo Qi, Yancheng Yuan
Abstract: Reinforcement learning from human feedback (RLHF) is an essential technique for ensuring that large language models (LLMs) are aligned with human values and preferences during the post-training phase. As an effective RLHF approach, group relative policy optimization (GRPO) has demonstrated success in many LLM-based applications. However, efficient GRPO-based RLHF training remains a challenge. Recent studies reveal that a higher reward variance of the initial policy model leads to faster RLHF training. Inspired by this finding, we propose a practical reward adjustment model to accelerate RLHF training by provably increasing the reward variance and preserving the relative preferences and reward expectation. Our reward adjustment method inherently poses a nonconvex optimization problem, which is NP-hard to solve in general. To overcome the computational challenges, we design a novel $O(n \log n)$ algorithm to find a global solution of the nonconvex reward adjustment model by explicitly characterizing the extreme points of the feasible set. As an important application, we naturally integrate this reward adjustment model into the GRPO algorithm, leading to a more efficient GRPO with reward variance increase (GRPOVI) algorithm for RLHF training. As an interesting byproduct, we provide an indirect explanation for the empirical effectiveness of GRPO with rule-based reward for RLHF training, as demonstrated in DeepSeek-R1. Experiment results demonstrate that the GRPOVI algorithm can significantly improve the RLHF training efficiency compared to the original GRPO algorithm.
Authors: Fangyikang Wang, Hubery Yin, Shaobin Zhuang, Huminhao Zhu, Yinan Li, Lei Qian, Chao Zhang, Hanbin Zhao, Hui Qian, Chen Li
Abstract: Recent Diffusion models (DMs) advancements have explored incorporating the second-order diffusion Fisher information (DF), defined as the negative Hessian of log density, into various downstream tasks and theoretical analysis. However, current practices typically approximate the diffusion Fisher by applying auto-differentiation to the learned score network. This black-box method, though straightforward, lacks any accuracy guarantee and is time-consuming. In this paper, we show that the diffusion Fisher actually resides within a space spanned by the outer products of score and initial data. Based on the outer-product structure, we develop two efficient approximation algorithms to access the trace and matrix-vector multiplication of DF, respectively. These algorithms bypass the auto-differentiation operations with time-efficient vector-product calculations. Furthermore, we establish the approximation error bounds for the proposed algorithms. Experiments in likelihood evaluation and adjoint optimization demonstrate the superior accuracy and reduced computational cost of our proposed algorithms. Additionally, based on the novel outer-product formulation of DF, we design the first numerical verification experiment for the optimal transport property of the general PF-ODE deduced map.
Authors: Haokun Chen, Yueqi Zhang, Yuan Bi, Yao Zhang, Tong Liu, Jinhe Bi, Jian Lan, Jindong Gu, Claudia Grosser, Denis Krompass, Nassir Navab, Volker Tresp
Abstract: In recent years, Large Language Models (LLMs) have achieved remarkable advancements, drawing significant attention from the research community. Their capabilities are largely attributed to large-scale architectures, which require extensive training on massive datasets. However, such datasets often contain sensitive or copyrighted content sourced from the public internet, raising concerns about data privacy and ownership. Regulatory frameworks, such as the General Data Protection Regulation (GDPR), grant individuals the right to request the removal of such sensitive information. This has motivated the development of machine unlearning algorithms that aim to remove specific knowledge from models without the need for costly retraining. Despite these advancements, evaluating the efficacy of unlearning algorithms remains a challenge due to the inherent complexity and generative nature of LLMs. In this work, we introduce a comprehensive auditing framework for unlearning evaluation, comprising three benchmark datasets, six unlearning algorithms, and five prompt-based auditing methods. By using various auditing algorithms, we evaluate the effectiveness and robustness of different unlearning strategies. To explore alternatives beyond prompt-based auditing, we propose a novel technique that leverages intermediate activation perturbations, addressing the limitations of auditing methods that rely solely on model inputs and outputs.
Authors: Muhammad Shafi, Syed Mohsin Bokhari
Abstract: Land cover and land use (LULC) changes are key applications of satellite imagery, and they have critical roles in resource management, urbanization, protection of soils and the environment, and enhancing sustainable development. The literature has heavily utilized multispectral spatiotemporal satellite data alongside advanced machine learning algorithms to monitor and predict LULC changes. This study analyzes and compares LULC changes across various governorates (provinces) of the Sultanate of Oman from 2016 to 2021 using annual time steps. For the chosen region, multispectral spatiotemporal data were acquired from the open-source Sentinel-2 satellite dataset. Supervised machine learning algorithms were used to train and classify different land covers, such as water bodies, crops, urban, etc. The constructed model was subsequently applied within the study region, allowing for an effective comparative evaluation of LULC changes within the given timeframe.
Authors: Yixin Ren, Chenghou Jin, Yewei Xia, Li Ke, Longtao Huang, Hui Xue, Hao Zhang, Jihong Guan, Shuigeng Zhou
Abstract: Determining conditional independence (CI) relationships between random variables is a fundamental yet challenging task in machine learning and statistics, especially in high-dimensional settings. Existing generative model-based CI testing methods, such as those utilizing generative adversarial networks (GANs), often struggle with undesirable modeling of conditional distributions and training instability, resulting in subpar performance. To address these issues, we propose a novel CI testing method via score-based generative modeling, which achieves precise Type I error control and strong testing power. Concretely, we first employ a sliced conditional score matching scheme to accurately estimate conditional score and use Langevin dynamics conditional sampling to generate null hypothesis samples, ensuring precise Type I error control. Then, we incorporate a goodness-of-fit stage into the method to verify generated samples and enhance interpretability in practice. We theoretically establish the error bound of conditional distributions modeled by score-based generative models and prove the validity of our CI tests. Extensive experiments on both synthetic and real-world datasets show that our method significantly outperforms existing state-of-the-art methods, providing a promising way to revitalize generative model-based CI testing.
Authors: Connor Cooper, Geoffrey I. Webb, Daniel F. Schmidt
Abstract: Bayesian network classifiers (BNCs) possess a number of properties desirable for a modern classifier: They are easily interpretable, highly scalable, and offer adaptable complexity. However, traditional methods for learning BNCs have historically underperformed when compared to leading classification methods such as random forests. Recent parameter smoothing techniques using hierarchical Dirichlet processes (HDPs) have enabled BNCs to achieve performance competitive with random forests on categorical data, but these techniques are relatively inflexible, and require a complicated, specialized sampling process. In this paper, we introduce a novel method for parameter estimation that uses a log-linear regression to approximate the behaviour of HDPs. As a linear model, our method is remarkably flexible and simple to interpret, and can leverage the vast literature on learning linear models. Our experiments show that our method can outperform HDP smoothing while being orders of magnitude faster, remaining competitive with random forests on categorical data.
Authors: Tu Bui, Mohamed Suliman, Aparajita Haldar, Mohammed Amer, Serban Georgescu
Abstract: Despite the transformative impact of deep learning on text, audio, and image datasets, its dominance in tabular data, especially in the medical domain where data are often scarce, remains less clear. In this paper, we propose X2Graph, a novel deep learning method that achieves strong performance on small biological tabular datasets. X2Graph leverages external knowledge about the relationships between table columns, such as gene interactions, to convert each sample into a graph structure. This transformation enables the application of standard message passing algorithms for graph modeling. Our X2Graph method demonstrates superior performance compared to existing tree-based and deep learning methods across three cancer subtyping datasets.
Authors: Chetan Verma, Aditya Srinivas Timmaraju, Cho Jui-Hsieh, Suyash Damle, Ngot Bui, Yang Zhang, Wen Chen, Xin Liu, Prateek Jain, Inderjit S Dhillon
Abstract: Industry-grade ML models are carefully designed to meet rapidly evolving serving constraints, which requires significant resources for model development. In this paper, we propose MatTA, a framework for training multiple accurate Student models using a novel Teacher-TA-Student recipe. TA models are larger versions of the Student models with higher capacity, and thus allow Student models to better relate to the Teacher model and also bring in more domain-specific expertise. Furthermore, multiple accurate Student models can be extracted from the TA model. Therefore, despite only one training run, our methodology provides multiple servable options to trade off accuracy for lower serving cost. We demonstrate the proposed method, MatTA, on proprietary datasets and models. Its practical efficacy is underscored by live A/B tests within a production ML system, demonstrating 20% improvement on a key metric. We also demonstrate our method on GPT-2 Medium, a public model, and achieve relative improvements of over 24% on SAT Math and over 10% on the LAMBADA benchmark.
Authors: Yang Liu, Deyu Bo, Wenxuan Cao, Yuan Fang, Yawen Li, Chuan Shi
Abstract: Graph self-supervised learning seeks to learn effective graph representations without relying on labeled data. Among various approaches, graph autoencoders (GAEs) have gained significant attention for their efficiency and scalability. Typically, GAEs take incomplete graphs as input and predict missing elements, such as masked nodes or edges. While effective, our experimental investigation reveals that traditional node or edge masking paradigms primarily capture low-frequency signals in the graph and fail to learn the expressive structural information. To address these issues, we propose Graph Positional Autoencoders (GraphPAE), which employs a dual-path architecture to reconstruct both node features and positions. Specifically, the feature path uses positional encoding to enhance the message-passing processing, improving GAE's ability to predict the corrupted information. The position path, on the other hand, leverages node representations to refine positions and approximate eigenvectors, thereby enabling the encoder to learn diverse frequency information. We conduct extensive experiments to verify the effectiveness of GraphPAE, including heterophilic node classification, graph property prediction, and transfer learning. The results demonstrate that GraphPAE achieves state-of-the-art performance and consistently outperforms baselines by a large margin.
Authors: Yuting Li, Shaoyuan Huang, Tengwen Zhang, Cheng Zhang, Xiaofei Wang, Victor C. M. Leung
Abstract: With the rapid growth of live streaming services, Crowdsourced Cloud-edge service Platforms (CCPs) are playing an increasingly important role in meeting the increasing demand. Although stream scheduling plays a critical role in optimizing CCPs' revenue, most optimization strategies struggle to achieve practical results due to various anomalies in unstable CCPs. Additionally, the substantial scale of CCPs magnifies the difficulties of anomaly detection in time-sensitive scheduling. To tackle these challenges, this paper proposes Sentinel, a proactive anomaly detection-based scheduling framework. Sentinel models the scheduling process as a two-stage Pre-Post-Scheduling paradigm: in the pre-scheduling stage, Sentinel conducts anomaly detection and constructs a strategy pool; in the post-scheduling stage, upon request arrival, it triggers an appropriate scheduling based on a pre-generated strategy to implement the scheduling process. Extensive experiments on realistic datasets show that Sentinel significantly reduces anomaly frequency by 70%, improves revenue by 74%, and doubles the scheduling speed.
Authors: Sheng Ouyang, Yulan Hu, Ge Chen, Qingyang Li, Fuzheng Zhang, Yong Liu
Abstract: Rewards serve as proxies for human preferences and play a crucial role in Reinforcement Learning from Human Feedback (RLHF). However, if these rewards are inherently imperfect, exhibiting various biases, they can adversely affect the alignment of large language models (LLMs). In this paper, we collectively define the various biases present in rewards as the problem of reward unfairness. We propose a bias-agnostic method to address the issue of reward fairness from a resource allocation perspective, without specifically designing for each type of bias, yet effectively mitigating them. Specifically, we model preference learning as a resource allocation problem, treating rewards as resources to be allocated while considering the trade-off between utility and fairness in their distribution. We propose two methods, Fairness Regularization and Fairness Coefficient, to achieve fairness in rewards. We apply our methods in both verification and reinforcement learning scenarios to obtain a fairness reward model and a policy model, respectively. Experiments conducted in these scenarios demonstrate that our approach aligns LLMs with human preferences in a more fair manner.
Authors: Maxiu Xiao, Jianglin Lan, Jingxing Yu, Eldert van Henten, Congcong Sun
Abstract: Climate control is crucial for greenhouse production as it directly affects crop growth and resource use. Reinforcement learning (RL) has received increasing attention in this field, but still faces challenges, including limited training efficiency and high reliance on initial learning conditions. Interactive RL, which combines human (grower) input with the RL agent's learning, offers a potential solution to overcome these challenges. However, interactive RL has not yet been applied to greenhouse climate control and may face challenges related to imperfect inputs. Therefore, this paper aims to explore the possibility and performance of applying interactive RL with imperfect inputs into greenhouse climate control, by: (1) developing three representative interactive RL algorithms tailored for greenhouse climate control (reward shaping, policy shaping and control sharing); (2) analyzing how input characteristics are often contradicting, and how the trade-offs between them make grower's inputs difficult to perfect; (3) proposing a neural network-based approach to enhance the robustness of interactive RL agents under limited input availability; (4) conducting a comprehensive evaluation of the three interactive RL algorithms with imperfect inputs in a simulated greenhouse environment. The demonstration shows that interactive RL incorporating imperfect grower inputs has the potential to improve the performance of the RL agent. RL algorithms that influence action selection, such as policy shaping and control sharing, perform better when dealing with imperfect inputs, achieving 8.4% and 6.8% improvement in profit, respectively. In contrast, reward shaping, an algorithm that manipulates the reward function, is sensitive to imperfect inputs and leads to a 9.4% decrease in profit. This highlights the importance of selecting an appropriate mechanism when incorporating imperfect inputs.
Authors: Mannmohan Muthuraman
Abstract: Dynamic Spectral Backpropagation (DSBP) enhances neural network training under resource constraints by projecting gradients onto principal eigenvectors, reducing complexity and promoting flat minima. Five extensions are proposed, dynamic spectral inference, spectral architecture optimization, spectral meta learning, spectral transfer regularization, and Lie algebra inspired dynamics, to address challenges in robustness, fewshot learning, and hardware efficiency. Supported by a third order stochastic differential equation (SDE) and a PAC Bayes limit, DSBP outperforms Sharpness Aware Minimization (SAM), Low Rank Adaptation (LoRA), and Model Agnostic Meta Learning (MAML) on CIFAR 10, Fashion MNIST, MedMNIST, and Tiny ImageNet, as demonstrated through extensive experiments and visualizations. Future work focuses on scalability, bias mitigation, and ethical considerations.
Authors: Roseline Polle, Agnes Norbury, Alexandra Livia Georgescu, Nicholas Cummins, Stefano Goria
Abstract: Speaker-dependent modelling can substantially improve performance in speech-based health monitoring applications. While mixed-effect models are commonly used for such speaker adaptation, they require computationally expensive retraining for each new observation, making them impractical in a production environment. We reformulate this task as a meta-learning problem and explore three approaches of increasing complexity: ensemble-based distance models, prototypical networks, and transformer-based sequence models. Using pre-trained speech embeddings, we evaluate these methods on a large longitudinal dataset of shift workers (N=1,185, 10,286 recordings), predicting time since sleep from speech as a function of fatigue, a symptom commonly associated with ill-health. Our results demonstrate that all meta-learning approaches tested outperformed both cross-sectional and conventional mixed-effects models, with a transformer-based method achieving the strongest performance.
Authors: Ahmad Anaqreh, Shih-Kai Chou, Mihael Mohor\v{c}i\v{c}, Carolina Fortuna
Abstract: Modeling propagation is the cornerstone for designing and optimizing next-generation wireless systems, with a particular emphasis on 5G and beyond era. Traditional modeling methods have long relied on statistic-based techniques to characterize propagation behavior across different environments. With the expansion of wireless communication systems, there is a growing demand for methods that guarantee the accuracy and interoperability of modeling. Artificial intelligence (AI)-based techniques, in particular, are increasingly being adopted to overcome this challenge, although the interpretability is not assured with most of these methods. Inspired by recent advancements in AI, this paper proposes a novel approach that accelerates the discovery of path loss models while maintaining interpretability. The proposed method automates the model formulation, evaluation, and refinement, facilitating model discovery. We evaluate two techniques: one based on Deep Symbolic Regression, offering full interpretability, and the second based on Kolmogorov-Arnold Networks, providing two levels of interpretability. Both approaches are evaluated on two synthetic and two real-world datasets. Our results show that Kolmogorov-Arnold Networks achieve R^2 values close to 1 with minimal prediction error, while Deep Symbolic Regression generates compact models with moderate accuracy. Moreover, on the selected examples, we demonstrate that automated methods outperform traditional methods, achieving up to 75% reduction in prediction errors, offering accurate and explainable solutions with potential to increase the efficiency of discovering next-generation path loss models.
Authors: Srishti Gupta, Daniele Angioni, Maura Pintor, Ambra Demontis, Lea Sch\"onherr, Battista Biggio, Fabio Roli
Abstract: Class-incremental learning (CIL) poses significant challenges in open-world scenarios, where models must not only learn new classes over time without forgetting previous ones but also handle inputs from unknown classes that a closed-set model would misclassify. Recent works address both issues by (i)~training multi-head models using the task-incremental learning framework, and (ii) predicting the task identity employing out-of-distribution (OOD) detectors. While effective, the latter mainly relies on joint training with a memory buffer of past data, raising concerns around privacy, scalability, and increased training time. In this paper, we present an in-depth analysis of post-hoc OOD detection methods and investigate their potential to eliminate the need for a memory buffer. We uncover that these methods, when applied appropriately at inference time, can serve as a strong substitute for buffer-based OOD detection. We show that this buffer-free approach achieves comparable or superior performance to buffer-based methods both in terms of class-incremental learning and the rejection of unknown samples. Experimental results on CIFAR-10, CIFAR-100 and Tiny ImageNet datasets support our findings, offering new insights into the design of efficient and privacy-preserving CIL systems for open-world settings.
Authors: Gaspard Oliviers, Mufeng Tang, Rafal Bogacz
Abstract: Predictive coding (PC) is an influential computational model of visual learning and inference in the brain. Classical PC was proposed as a top-down generative model, where the brain actively predicts upcoming visual inputs, and inference minimises the prediction errors. Recent studies have also shown that PC can be formulated as a discriminative model, where sensory inputs predict neural activities in a feedforward manner. However, experimental evidence suggests that the brain employs both generative and discriminative inference, while unidirectional PC models show degraded performance in tasks requiring bidirectional processing. In this work, we propose bidirectional PC (bPC), a PC model that incorporates both generative and discriminative inference while maintaining a biologically plausible circuit implementation. We show that bPC matches or outperforms unidirectional models in their specialised generative or discriminative tasks, by developing an energy landscape that simultaneously suits both tasks. We also demonstrate bPC's superior performance in two biologically relevant tasks including multimodal learning and inference with missing information, suggesting that bPC resembles biological visual inference more closely.
Authors: Zheming Zhang, Yan Jiang, Qingshan Li, Ai Han
Abstract: In China's competitive fresh e-commerce market, optimizing operational strategies, especially inventory management in front-end warehouses, is key to enhance customer satisfaction and to gain a competitive edge. Front-end warehouses are placed in residential areas to ensure the timely delivery of fresh goods and are usually in small size. This brings the challenge of deciding which goods to stock and in what quantities, taking into account capacity constraints. To address this issue, traditional predict-then-optimize (PTO) methods that predict sales and then decide on inventory often don't align prediction with inventory goals, as well as fail to prioritize consumer satisfaction. This paper proposes a multi-task Optimize-then-Predict-then-Optimize (OTPTO) approach that jointly optimizes product selection and inventory management, aiming to increase consumer satisfaction by maximizing the full order fulfillment rate. Our method employs a 0-1 mixed integer programming model OM1 to determine historically optimal inventory levels, and then uses a product selection model PM1 and the stocking model PM2 for prediction. The combined results are further refined through a post-processing algorithm OM2. Experimental results from JD.com's 7Fresh platform demonstrate the robustness and significant advantages of our OTPTO method. Compared to the PTO approach, our OTPTO method substantially enhances the full order fulfillment rate by 4.34% (a relative increase of 7.05%) and narrows the gap to the optimal full order fulfillment rate by 5.27%. These findings substantiate the efficacy of the OTPTO method in managing inventory at front-end warehouses of fresh e-commerce platforms and provide valuable insights for future research in this domain.
Authors: Yinuo Wang, Mining Tan, Wenjun Zou, Haotian Lin, Xujie Song, Wenxuan Wang, Tong Liu, Likun Wang, Guojian Zhan, Tianze Zhu, Shiqi Liu, Jingliang Duan, Shengbo Eben Li
Abstract: Due to their expressive capacity, diffusion models have shown great promise in offline RL and imitation learning. Diffusion Actor-Critic with Entropy Regulator (DACER) extended this capability to online RL by using the reverse diffusion process as a policy approximator, trained end-to-end with policy gradient methods, achieving strong performance. However, this comes at the cost of requiring many diffusion steps, which significantly hampers training efficiency, while directly reducing the steps leads to noticeable performance degradation. Critically, the lack of inference efficiency becomes a significant bottleneck for applying diffusion policies in real-time online RL settings. To improve training and inference efficiency while maintaining or even enhancing performance, we propose a Q-gradient field objective as an auxiliary optimization target to guide the denoising process at each diffusion step. Nonetheless, we observe that the independence of the Q-gradient field from the diffusion time step negatively impacts the performance of the diffusion policy. To address this, we introduce a temporal weighting mechanism that enables the model to efficiently eliminate large-scale noise in the early stages and refine actions in the later stages. Experimental results on MuJoCo benchmarks and several multimodal tasks demonstrate that the DACER2 algorithm achieves state-of-the-art performance in most MuJoCo control tasks with only five diffusion steps, while also exhibiting stronger multimodality compared to DACER.
Authors: Monika Gahalawat, Maneesh Bilalpur, Raul Fernandez Rojas, Jeffrey F. Cohn, Roland Goecke, Ramanathan Subramanian
Abstract: Depression is a debilitating mood disorder negatively impacting millions worldwide. While researchers have explored multiple verbal and non-verbal behavioural cues for automated depression assessment, head motion has received little attention thus far. Further, the common practice of validating machine learning models via a single dataset can limit model generalisability. This work examines the effectiveness and generalisability of models utilising elementary head motion units, termed kinemes, for depression severity estimation. Specifically, we consider three depression datasets from different western cultures (German: AVEC2013, Australian: Blackdog and American: Pitt datasets) with varied contextual and recording settings to investigate the generalisability of the derived kineme patterns via two methods: (i) k-fold cross-validation over individual/multiple datasets, and (ii) model reuse on other datasets. Evaluating classification and regression performance with classical machine learning methods, our results show that: (1) head motion patterns are efficient biomarkers for estimating depression severity, achieving highly competitive performance for both classification and regression tasks on a variety of datasets, including achieving the second best Mean Absolute Error (MAE) on the AVEC2013 dataset, and (2) kineme-based features are more generalisable than (a) raw head motion descriptors for binary severity classification, and (b) other visual behavioural cues for severity estimation (regression).
Authors: Jian Yao, Ran Cheng, Xingyu Wu, Jibin Wu, Kay Chen Tan
Abstract: The reasoning capabilities of large language models (LLMs) have advanced rapidly, particularly following the release of DeepSeek R1, which has inspired a surge of research into data quality and reinforcement learning (RL) algorithms. Despite the pivotal role diversity plays in RL, its influence on LLM reasoning remains largely underexplored. To bridge this gap, this work presents a systematic investigation into the impact of diversity in RL-based training for LLM reasoning, and proposes a novel diversity-aware policy optimization method. Across evaluations on 12 LLMs, we observe a strong positive correlation between the solution diversity and Potential at k (a novel metric quantifying an LLM's reasoning potential) in high-performing models. This finding motivates our method to explicitly promote diversity during RL training. Specifically, we design a token-level diversity and reformulate it into a practical objective, then we selectively apply it to positive samples. Integrated into the R1-zero training framework, our method achieves a 3.5 percent average improvement across four mathematical reasoning benchmarks, while generating more diverse and robust solutions.
Authors: Antonio Ferrara, Andrea Pugnana, Francesco Bonchi, Salvatore Ruggieri
Abstract: Ranking systems influence decision-making in high-stakes domains like health, education, and employment, where they can have substantial economic and social impacts. This makes the integration of safety mechanisms essential. One such mechanism is $\textit{abstention}$, which enables algorithmic decision-making system to defer uncertain or low-confidence decisions to human experts. While abstention have been predominantly explored in the context of classification tasks, its application to other machine learning paradigms remains underexplored. In this paper, we introduce a novel method for abstention in pairwise learning-to-rank tasks. Our approach is based on thresholding the ranker's conditional risk: the system abstains from making a decision when the estimated risk exceeds a predefined threshold. Our contributions are threefold: a theoretical characterization of the optimal abstention strategy, a model-agnostic, plug-in algorithm for constructing abstaining ranking models, and a comprehensive empirical evaluations across multiple datasets, demonstrating the effectiveness of our approach.
Authors: Linyu Li, Zhi Jin, Yuanpeng He, Dongming Jin, Haoran Duan, Zhengwei Tao, Xuan Zhang, Jiandong Li
Abstract: Knowledge graph completion (KGC) has attracted considerable attention in recent years because it is critical to improving the quality of knowledge graphs. Researchers have continuously explored various models. However, most previous efforts have neglected to take advantage of regularization from a deeper perspective and therefore have not been used to their full potential. This paper rethinks the application of regularization methods in KGC. Through extensive empirical studies on various KGC models, we find that carefully designed regularization not only alleviates overfitting and reduces variance but also enables these models to break through the upper bounds of their original performance. Furthermore, we introduce a novel sparse-regularization method that embeds the concept of rank-based selective sparsity into the KGC regularizer. The core idea is to selectively penalize those components with significant features in the embedding vector, thus effectively ignoring many components that contribute little and may only represent noise. Various comparative experiments on multiple datasets and multiple models show that the SPR regularization method is better than other regularization methods and can enable the KGC model to further break through the performance margin.
Authors: Benyamin Trachtenberg, Nir Rosenfeld
Abstract: In strategic classification, the standard supervised learning setting is extended to support the notion of strategic user behavior in the form of costly feature manipulations made in response to a classifier. While standard learning supports a broad range of model classes, the study of strategic classification has, so far, been dedicated mostly to linear classifiers. This work aims to expand the horizon by exploring how strategic behavior manifests under non-linear classifiers and what this implies for learning. We take a bottom-up approach showing how non-linearity affects decision boundary points, classifier expressivity, and model classes complexity. A key finding is that universal approximators (e.g., neural nets) are no longer universal once the environment is strategic. We demonstrate empirically how this can create performance gaps even on an unrestricted model class.
Authors: Pirzada Suhail, Rehna Afroz, Amit Sethi
Abstract: Out-of-distribution (OOD) detection and uncertainty estimation (UE) are critical components for building safe machine learning systems, especially in real-world scenarios where unexpected inputs are inevitable. In this work, we propose a novel framework that combines network inversion with classifier training to simultaneously address both OOD detection and uncertainty estimation. For a standard n-class classification task, we extend the classifier to an (n+1)-class model by introducing a "garbage" class, initially populated with random gaussian noise to represent outlier inputs. After each training epoch, we use network inversion to reconstruct input images corresponding to all output classes that initially appear as noisy and incoherent and are therefore excluded to the garbage class for retraining the classifier. This cycle of training, inversion, and exclusion continues iteratively till the inverted samples begin to resemble the in-distribution data more closely, suggesting that the classifier has learned to carve out meaningful decision boundaries while sanitising the class manifolds by pushing OOD content into the garbage class. During inference, this training scheme enables the model to effectively detect and reject OOD samples by classifying them into the garbage class. Furthermore, the confidence scores associated with each prediction can be used to estimate uncertainty for both in-distribution and OOD inputs. Our approach is scalable, interpretable, and does not require access to external OOD datasets or post-hoc calibration techniques while providing a unified solution to the dual challenges of OOD detection and uncertainty estimation.
Authors: Kevin Frans, Seohong Park, Pieter Abbeel, Sergey Levine
Abstract: At the core of reinforcement learning is the idea of learning beyond the performance in the data. However, scaling such systems has proven notoriously tricky. In contrast, techniques from generative modeling have proven remarkably scalable and are simple to train. In this work, we combine these strengths, by deriving a direct relation between policy improvement and guidance of diffusion models. The resulting framework, CFGRL, is trained with the simplicity of supervised learning, yet can further improve on the policies in the data. On offline RL tasks, we observe a reliable trend -- increased guidance weighting leads to increased performance. Of particular importance, CFGRL can operate without explicitly learning a value function, allowing us to generalize simple supervised methods (e.g., goal-conditioned behavioral cloning) to further prioritize optimality, gaining performance for "free" across the board.
Authors: Safwan Labbi, Paul Mangold, Daniil Tiapkin, Eric Moulines
Abstract: Ensuring convergence of policy gradient methods in federated reinforcement learning (FRL) under environment heterogeneity remains a major challenge. In this work, we first establish that heterogeneity, perhaps counter-intuitively, can necessitate optimal policies to be non-deterministic or even time-varying, even in tabular environments. Subsequently, we prove global convergence results for federated policy gradient (FedPG) algorithms employing local updates, under a {\L}ojasiewicz condition that holds only for each individual agent, in both entropy-regularized and non-regularized scenarios. Crucially, our theoretical analysis shows that FedPG attains linear speed-up with respect to the number of agents, a property central to efficient federated learning. Leveraging insights from our theoretical findings, we introduce b-RS-FedPG, a novel policy gradient method that employs a carefully constructed softmax-inspired parameterization coupled with an appropriate regularization scheme. We further demonstrate explicit convergence rates for b-RS-FedPG toward near-optimal stationary policies. Finally, we demonstrate that empirically both FedPG and b-RS-FedPG consistently outperform federated Q-learning on heterogeneous settings.
Authors: Chenjie Li, Amir Gilad, Boris Glavic, Zhengjie Miao, Sudeepa Roy
Abstract: Programmatic weak supervision (PWS) significantly reduces human effort for labeling data by combining the outputs of user-provided labeling functions (LFs) on unlabeled datapoints. However, the quality of the generated labels depends directly on the accuracy of the LFs. In this work, we study the problem of fixing LFs based on a small set of labeled examples. Towards this goal, we develop novel techniques for repairing a set of LFs by minimally changing their results on the labeled examples such that the fixed LFs ensure that (i) there is sufficient evidence for the correct label of each labeled datapoint and (ii) the accuracy of each repaired LF is sufficiently high. We model LFs as conditional rules which enables us to refine them, i.e., to selectively change their output for some inputs. We demonstrate experimentally that our system improves the quality of LFs based on surprisingly small sets of labeled datapoints.
Authors: Ildus Sadrtdinov, Ivan Klimov, Ekaterina Lobacheva, Dmitry Vetrov
Abstract: We present a thermodynamic interpretation of the stationary behavior of stochastic gradient descent (SGD) under fixed learning rates (LRs) in neural network training. We show that SGD implicitly minimizes a free energy function $F=U-TS$, balancing training loss $U$ and the entropy of the weights distribution $S$, with temperature $T$ determined by the LR. This perspective offers a new lens on why high LRs prevent training from converging to the loss minima and how different LRs lead to stabilization at different loss levels. We empirically validate the free energy framework on both underparameterized (UP) and overparameterized (OP) models. UP models consistently follow free energy minimization, with temperature increasing monotonically with LR, while for OP models, the temperature effectively drops to zero at low LRs, causing SGD to minimize the loss directly and converge to an optimum. We attribute this mismatch to differences in the signal-to-noise ratio of stochastic gradients near optima, supported by both a toy example and neural network experiments.
Authors: Sabina J. Sloman, Michele Caprio, Samuel Kaski
Abstract: When data are noisy, a statistical learner's goal is to resolve epistemic uncertainty about the data it will encounter at test-time, i.e., to identify the distribution of test (target) data. Many real-world learning settings introduce sources of epistemic uncertainty that can not be resolved on the basis of training (source) data alone: The source data may arise from multiple tasks (multitask learning), the target data may differ systematically from the source data tasks (distribution shift), and/or the learner may not arrive at an accurate characterization of the source data (imperfect learning). We introduce a principled definition of epistemic error, and provide a generic, decompositional epistemic error bound. Our error bound is the first to (i) consider epistemic error specifically, (ii) accommodate all the sources of epistemic uncertainty above, and (iii) separately attribute the error to each of multiple aspects of the learning procedure and environment. As corollaries of the generic result, we provide (i) epistemic error bounds specialized to the settings of Bayesian transfer learning and distribution shift within $\epsilon$-neighborhoods, and (ii) a set of corresponding generalization bounds. Finally, we provide a novel definition of negative transfer, and validate its insights in a synthetic experimental setting.
Authors: Sebasti\'an Jim\'enez, Mira J\"urgens, Willem Waegeman
Abstract: In recent years various supervised learning methods that disentangle aleatoric and epistemic uncertainty based on second-order distributions have been proposed. We argue that these methods fail to capture critical components of epistemic uncertainty, particularly due to the often-neglected component of model bias. To show this, we make use of a more fine-grained taxonomy of epistemic uncertainty sources in machine learning models, and analyse how the classical bias-variance decomposition of the expected prediction error can be decomposed into different parts reflecting these uncertainties. By using a simulation-based evaluation protocol which encompasses epistemic uncertainty due to both procedural- and data-driven uncertainty components, we illustrate that current methods rarely capture the full spectrum of epistemic uncertainty. Through theoretical insights and synthetic experiments, we show that high model bias can lead to misleadingly low estimates of epistemic uncertainty, and common second-order uncertainty quantification methods systematically blur bias-induced errors into aleatoric estimates, thereby underrepresenting epistemic uncertainty. Our findings underscore that meaningful aleatoric estimates are feasible only if all relevant sources of epistemic uncertainty are properly represented.
Authors: Yu Zhang, Dong Guo, Fang Wu, Guoliang Zhu, Dian Ding, Yiming Zhang
Abstract: Large Language Models (LLMs) with extended context lengths face significant computational challenges during the pre-filling phase, primarily due to the quadratic complexity of self-attention. Existing methods typically employ dynamic pattern matching and block-sparse low-level implementations. However, their reliance on local information for pattern identification fails to capture global contexts, and the coarse granularity of blocks leads to persistent internal sparsity, resulting in suboptimal accuracy and efficiency. To address these limitations, we propose \textbf{AnchorAttention}, a difference-aware, dynamic sparse attention mechanism that efficiently identifies critical attention regions at a finer stripe granularity while adapting to global contextual information, achieving superior speed and accuracy. AnchorAttention comprises three key components: (1) \textbf{Pattern-based Anchor Computation}, leveraging the commonalities present across all inputs to rapidly compute a set of near-maximum scores as the anchor; (2) \textbf{Difference-aware Stripe Sparsity Identification}, performing difference-aware comparisons with the anchor to quickly obtain discrete coordinates of significant regions in a stripe-like sparsity pattern; (3) \textbf{Fine-grained Sparse Computation}, replacing the traditional contiguous KV block loading approach with simultaneous discrete KV position loading to maximize sparsity rates while preserving full hardware computational potential. With its finer-grained sparsity strategy, \textbf{AnchorAttention} achieves higher sparsity rates at the same recall level, significantly reducing computation time. Compared to previous state-of-the-art methods, at a text length of 128k, it achieves a speedup of 1.44$\times$ while maintaining higher recall rates.
Authors: Arjun Devraj, Eric Ding, Abhishek Vijaya Kumar, Robert Kleinberg, Rachee Singh
Abstract: Distributed machine learning workloads use data and tensor parallelism for training and inference, both of which rely on the AllReduce collective to synchronize gradients or activations. However, bulk-synchronous AllReduce algorithms can be delayed by a persistent straggler that is slower to reach the synchronization barrier required to begin the collective. To address this challenge, we propose StragglAR: an AllReduce algorithm that accelerates distributed training and inference in the presence of persistent stragglers. StragglAR implements a ReduceScatter among the remaining GPUs during the straggler-induced delay, and then executes a novel collective algorithm to complete the AllReduce once the straggler reaches the synchronization barrier. StragglAR achieves a 2x theoretical speedup over popular bandwidth-efficient AllReduce algorithms (e.g., Ring) for large GPU clusters with persistent stragglers. On an 8-GPU server, our implementation of StragglAR yields a 22% speedup over state-of-the-art AllReduce algorithms.
Authors: Raj Ghugare, Benjamin Eysenbach
Abstract: Modern reinforcement learning (RL) algorithms have found success by using powerful probabilistic models, such as transformers, energy-based models, and diffusion/flow-based models. To this end, RL researchers often choose to pay the price of accommodating these models into their algorithms -- diffusion models are expressive, but are computationally intensive due to their reliance on solving differential equations, while autoregressive transformer models are scalable but typically require learning discrete representations. Normalizing flows (NFs), by contrast, seem to provide an appealing alternative, as they enable likelihoods and sampling without solving differential equations or autoregressive architectures. However, their potential in RL has received limited attention, partly due to the prevailing belief that normalizing flows lack sufficient expressivity. We show that this is not the case. Building on recent work in NFs, we propose a single NF architecture which integrates seamlessly into RL algorithms, serving as a policy, Q-function, and occupancy measure. Our approach leads to much simpler algorithms, and achieves higher performance in imitation learning, offline, goal conditioned RL and unsupervised RL.
Authors: Maria Eleftheria Vlontzou, Maria Athanasiou, Christos Davatzikos, Konstantina S. Nikita
Abstract: The present study performs a comprehensive fairness analysis of machine learning (ML) models for the diagnosis of Mild Cognitive Impairment (MCI) and Alzheimer's disease (AD) from MRI-derived neuroimaging features. Biases associated with age, race, and gender in a multi-cohort dataset, as well as the influence of proxy features encoding these sensitive attributes, are investigated. The reliability of various fairness definitions and metrics in the identification of such biases is also assessed. Based on the most appropriate fairness measures, a comparative analysis of widely used pre-processing, in-processing, and post-processing bias mitigation strategies is performed. Moreover, a novel composite measure is introduced to quantify the trade-off between fairness and performance by considering the F1-score and the equalized odds ratio, making it appropriate for medical diagnostic applications. The obtained results reveal the existence of biases related to age and race, while no significant gender bias is observed. The deployed mitigation strategies yield varying improvements in terms of fairness across the different sensitive attributes and studied subproblems. For race and gender, Reject Option Classification improves equalized odds by 46% and 57%, respectively, and achieves harmonic mean scores of 0.75 and 0.80 in the MCI versus AD subproblem, whereas for age, in the same subproblem, adversarial debiasing yields the highest equalized odds improvement of 40% with a harmonic mean score of 0.69. Insights are provided into how variations in AD neuropathology and risk factors, associated with demographic characteristics, influence model fairness.
Authors: Shifeng Xie, Aref Einizade, Jhony H. Giraldo
Abstract: Graph Representation Learning (GRL) is a fundamental task in machine learning, aiming to encode high-dimensional graph-structured data into low-dimensional vectors. Self-Supervised Learning (SSL) methods are widely used in GRL because they can avoid expensive human annotation. In this work, we propose a novel Subgraph Gaussian Embedding Contrast (SubGEC) method. Our approach introduces a subgraph Gaussian embedding module, which adaptively maps subgraphs to a structured Gaussian space, ensuring the preservation of input subgraph characteristics while generating subgraphs with a controlled distribution. We then employ optimal transport distances, more precisely the Wasserstein and Gromov-Wasserstein distances, to effectively measure the similarity between subgraphs, enhancing the robustness of the contrastive learning process. Extensive experiments across multiple benchmarks demonstrate that \method~outperforms or presents competitive performance against state-of-the-art approaches. Our findings provide insights into the design of SSL methods for GRL, emphasizing the importance of the distribution of the generated contrastive pairs.
Authors: Giorgos Iacovides, Wuyang Zhou, Chao Li, Qibin Zhao, Danilo Mandic
Abstract: Tensor networks (TNs) provide efficient representations of high-dimensional data, yet identification of the optimal TN structures, the so called tensor network structure search (TN-SS) problem, remains a challenge. Current state-of-the-art (SOTA) algorithms are computationally expensive as they require extensive function evaluations, which is prohibitive for real-world applications. In addition, existing methods ignore valuable domain information inherent in real-world tensor data and lack transparency in their identified TN structures. To this end, we propose a novel TN-SS framework, termed the tnLLM, which incorporates domain information about the data and harnesses the reasoning capabilities of large language models (LLMs) to directly predict suitable TN structures. The proposed framework involves a domain-aware prompting pipeline which instructs the LLM to infer suitable TN structures based on the real-world relationships between tensor modes. In this way, our approach is capable of not only iteratively optimizing the objective function, but also generating domain-aware explanations for the identified structures. Experimental results demonstrate that tnLLM achieves comparable TN-SS objective function values with much fewer function evaluations compared to SOTA algorithms. Furthermore, we demonstrate that the LLM-enabled domain information can be used to find good initializations in the search space for sampling-based SOTA methods to accelerate their convergence while preserving theoretical performance guarantees.
Authors: Alex Adams
Abstract: This paper investigates the comparative performance of two fundamental approaches to solving linear regression problems: the closed-form Moore-Penrose pseudoinverse and the iterative gradient descent method. Linear regression is a cornerstone of predictive modeling, and the choice of solver can significantly impact efficiency and accuracy. I review and discuss the theoretical underpinnings of both methods, analyze their computational complexity, and evaluate their empirical behavior on synthetic datasets with controlled characteristics, as well as on established real-world datasets. My results delineate the conditions under which each method excels in terms of computational time, numerical stability, and predictive accuracy. This work aims to provide practical guidance for researchers and practitioners in machine learning when selecting between direct, exact solutions and iterative, approximate solutions for linear regression tasks.
Authors: Yanzhao Hou, Jiaxiang Geng, Boyu Li, Xiaofeng Tao, Juncheng Wang, Xiaodong Xu, Bing Luo
Abstract: Federated LoRA has emerged as a promising technique for efficiently fine-tuning large language models (LLMs) on distributed devices by reducing the number of trainable parameters. However, existing approaches often inadequately overlook the theoretical and practical implications of system and data heterogeneity, thereby failing to optimize the overall training efficiency, particularly in terms of wall-clock time. In this paper, we propose an adaptive federated LoRA strategy with independent client sampling to minimize the convergence wall-clock time of federated fine-tuning under both computation and communication heterogeneity. We first derive a new convergence bound for federated LoRA with arbitrary and independent client sampling, notably without requiring the stringent bounded gradient assumption. Then, we introduce an adaptive bandwidth allocation scheme that accounts for heterogeneous client resources and system bandwidth constraints. Based on the derived theory, we formulate and solve a non-convex optimization problem to jointly determine the LoRA sketching ratios and sampling probabilities, aiming to minimize wall-clock convergence time. An efficient and low-complexity algorithm is developed to approximate the solution. Finally, extensive experiments demonstrate that our approach significantly reduces wall-clock training time compared to state-of-the-art methods across various models and datasets.
Authors: Yiran Guo, Lijie Xu, Jie Liu, Dan Ye, Shuang Qiu
Abstract: Enhancing the reasoning capabilities of large language models effectively using reinforcement learning (RL) remains a crucial challenge. Existing approaches primarily adopt two contrasting advantage estimation granularities: Token-level methods (e.g., PPO) aim to provide the fine-grained advantage signals but suffer from inaccurate estimation due to difficulties in training an accurate critic model. On the other extreme, trajectory-level methods (e.g., GRPO) solely rely on a coarse-grained advantage signal from the final reward, leading to imprecise credit assignment. To address these limitations, we propose Segment Policy Optimization (SPO), a novel RL framework that leverages segment-level advantage estimation at an intermediate granularity, achieving a better balance by offering more precise credit assignment than trajectory-level methods and requiring fewer estimation points than token-level methods, enabling accurate advantage estimation based on Monte Carlo (MC) without a critic model. SPO features three components with novel strategies: (1) flexible segment partition; (2) accurate segment advantage estimation; and (3) policy optimization using segment advantages, including a novel probability-mask strategy. We further instantiate SPO for two specific scenarios: (1) SPO-chain for short chain-of-thought (CoT), featuring novel cutpoint-based partition and chain-based advantage estimation, achieving $6$-$12$ percentage point improvements in accuracy over PPO and GRPO on GSM8K. (2) SPO-tree for long CoT, featuring novel tree-based advantage estimation, which significantly reduces the cost of MC estimation, achieving $7$-$11$ percentage point improvements over GRPO on MATH500 under 2K and 4K context evaluation. We make our code publicly available at https://github.com/AIFrameResearch/SPO.
Authors: Jiashuo Liu, Tianyu Wang, Henry Lam, Hongseok Namkoong, Jose Blanchet
Abstract: We introduce dro, an open-source Python library for distributionally robust optimization (DRO) for regression and classification problems. The library implements 14 DRO formulations and 9 backbone models, enabling 79 distinct DRO methods. Furthermore, dro is compatible with both scikit-learn and PyTorch. Through vectorization and optimization approximation techniques, dro reduces runtime by 10x to over 1000x compared to baseline implementations on large-scale datasets. Comprehensive documentation is available at https://python-dro.org.
URLs: https://python-dro.org.
Authors: Samo Hromadka, Kai Biegun, Lior Fox, James Heald, Maneesh Sahani
Abstract: We introduce a novel unsupervised learning method for time series data with latent dynamical structure: the recognition-parametrized Gaussian state space model (RP-GSSM). The RP-GSSM is a probabilistic model that learns Markovian Gaussian latents explaining statistical dependence between observations at different time steps, combining the intuition of contrastive methods with the flexible tools of probabilistic generative models. Unlike contrastive approaches, the RP-GSSM is a valid probabilistic model learned via maximum likelihood. Unlike generative approaches, the RP-GSSM has no need for an explicit network mapping from latents to observations, allowing it to focus model capacity on inference of latents. The model is both tractable and expressive: it admits exact inference thanks to its jointly Gaussian latent prior, while maintaining expressivity with an arbitrarily nonlinear neural network link between observations and latents. These qualities allow the RP-GSSM to learn task-relevant latents without ad-hoc regularization, auxiliary losses, or optimizer scheduling. We show how this approach outperforms alternatives on problems that include learning nonlinear stochastic dynamics from video, with or without background distractors. Our results position the RP-GSSM as a useful foundation model for a variety of downstream applications.
Authors: Adibvafa Fallahpour, Andrew Magnuson, Purav Gupta, Shihao Ma, Jack Naimer, Arnav Shah, Haonan Duan, Omar Ibrahim, Hani Goodarzi, Chris J. Maddison, Bo Wang
Abstract: Unlocking deep, interpretable biological reasoning from complex genomic data is a major AI challenge hindering scientific discovery. Current DNA foundation models, despite strong sequence representation, struggle with multi-step reasoning and lack inherent transparent, biologically intuitive explanations. We introduce BioReason, a pioneering architecture that, for the first time, deeply integrates a DNA foundation model with a Large Language Model (LLM). This novel connection enables the LLM to directly process and reason with genomic information as a fundamental input, fostering a new form of multimodal biological understanding. BioReason's sophisticated multi-step reasoning is developed through supervised fine-tuning and targeted reinforcement learning, guiding the system to generate logical, biologically coherent deductions. On biological reasoning benchmarks including KEGG-based disease pathway prediction - where accuracy improves from 88% to 97% - and variant effect prediction, BioReason demonstrates an average 15% performance gain over strong single-modality baselines. BioReason reasons over unseen biological entities and articulates decision-making through interpretable, step-by-step biological traces, offering a transformative approach for AI in biology that enables deeper mechanistic insights and accelerates testable hypothesis generation from genomic data. Data, code, and checkpoints are publicly available at https://github.com/bowang-lab/BioReason
Authors: Zhiding Liu, Mingyue Cheng, Guanhao Zhao, Jiqian Yang, Qi Liu, Enhong Chen
Abstract: Time series forecasting plays a vital role in various real-world applications and has attracted significant attention in recent decades. While recent methods have achieved remarkable accuracy by incorporating advanced inductive biases and training strategies, we observe that instance-level variations remain a significant challenge. These variations--stemming from distribution shifts, missing data, and long-tail patterns--often lead to suboptimal forecasts for specific instances, even when overall performance appears strong. To address this issue, we propose a model-agnostic framework, PIR, designed to enhance forecasting performance through Post-forecasting Identification and Revision. Specifically, PIR first identifies biased forecasting instances by estimating their accuracy. Based on this, the framework revises the forecasts using contextual information, including covariates and historical time series, from both local and global perspectives in a post-processing fashion. Extensive experiments on real-world datasets with mainstream forecasting models demonstrate that PIR effectively mitigates instance-level errors and significantly improves forecasting reliability.
Authors: Yaru Hao, Li Dong, Xun Wu, Shaohan Huang, Zewen Chi, Furu Wei
Abstract: Reinforcement learning algorithms are fundamental to align large language models with human preferences and to enhance their reasoning capabilities. However, current reinforcement learning algorithms often suffer from training instability due to loose on-policy constraints and computational inefficiency due to auxiliary models. In this work, we propose On-Policy RL with Optimal reward baseline (OPO), a novel and simplified reinforcement learning algorithm designed to address these challenges. OPO emphasizes the importance of exact on-policy training, which empirically stabilizes the training process and enhances exploration. Moreover, OPO introduces the optimal reward baseline that theoretically minimizes gradient variance. We evaluate OPO on mathematical reasoning benchmarks. The results demonstrate its superior performance and training stability without additional models or regularization terms. Furthermore, OPO achieves lower policy shifts and higher output entropy, encouraging more diverse and less repetitive responses. These results highlight OPO as a promising direction for stable and effective reinforcement learning in large language model alignment and reasoning tasks. The implementation is provided at https://github.com/microsoft/LMOps/tree/main/opo.
Authors: Mrinmay Sen, Sidhant R Nair, C Krishna Mohan
Abstract: This paper explores second-order optimization methods in Federated Learning (FL), addressing the critical challenges of slow convergence and the excessive communication rounds required to achieve optimal performance from the global model. While existing surveys in FL primarily focus on challenges related to statistical and device label heterogeneity, as well as privacy and security concerns in first-order FL methods, less attention has been given to the issue of slow model training. This slow training often leads to the need for excessive communication rounds or increased communication costs, particularly when data across clients are highly heterogeneous. In this paper, we examine various FL methods that leverage second-order optimization to accelerate the training process. We provide a comprehensive categorization of state-of-the-art second-order FL methods and compare their performance based on convergence speed, computational cost, memory usage, transmission overhead, and generalization of the global model. Our findings show the potential of incorporating Hessian curvature through second-order optimization into FL and highlight key challenges, such as the efficient utilization of Hessian and its inverse in FL. This work lays the groundwork for future research aimed at developing scalable and efficient federated optimization methods for improving the training of the global model in FL.
Authors: Nikita Agrawal, Simon Mertel, Ruben Mayer
Abstract: Post-training of foundation language models has emerged as a promising research domain in federated learning (FL) with the goal to enable privacy-preserving model improvements and adaptations to user's downstream tasks. Recent advances in this area adopt centralized post-training approaches that build upon black-box foundation language models where there is no access to model weights and architecture details. Although the use of black-box models has been successful in centralized post-training, their blind replication in FL raises several concerns. Our position is that using black-box models in FL contradicts the core principles of federation such as data privacy and autonomy. In this position paper, we critically analyze the usage of black-box models in federated post-training, and provide a detailed account of various aspects of openness and their implications for FL.
Authors: Radzim Sendyka, Christian Cabrera, Andrei Paleyes, Diana Robinson, Neil Lawrence
Abstract: This paper investigates the ability of large language models (LLMs) to recognise and solve tasks which have been obfuscated beyond recognition. Focusing on competitive programming and benchmark tasks (LeetCode and MATH), we compare performance across multiple models and obfuscation methods, such as noise and redaction. We demonstrate that all evaluated LLMs can solve tasks obfuscated to a level where the text would be unintelligible to human readers, and does not contain key pieces of instruction or context. We introduce the concept of eager pattern matching to describe this behaviour, which is not observed in tasks published after the models' knowledge cutoff date, indicating strong memorisation or overfitting to training data, rather than legitimate reasoning about the presented problem. We report empirical evidence of distinct performance decay patterns between contaminated and unseen datasets. We discuss the implications for benchmarking and evaluations of model behaviour, arguing for caution when designing experiments using standard datasets. We also propose measuring the decay of performance under obfuscation as a possible strategy for detecting dataset contamination and highlighting potential safety risks and interpretability issues for automated software systems.
Authors: Eitan Levin, Yuxin Ma, Mateo D\'iaz, Soledad Villar
Abstract: Many modern learning tasks require models that can take inputs of varying sizes. Consequently, dimension-independent architectures have been proposed for domains where the inputs are graphs, sets, and point clouds. Recent work on graph neural networks has explored whether a model trained on low-dimensional data can transfer its performance to higher-dimensional inputs. We extend this body of work by introducing a general framework for transferability across dimensions. We show that transferability corresponds precisely to continuity in a limit space formed by identifying small problem instances with equivalent large ones. This identification is driven by the data and the learning task. We instantiate our framework on existing architectures, and implement the necessary changes to ensure their transferability. Finally, we provide design principles for designing new transferable models. Numerical experiments support our findings.
Authors: Qingyu Shi, Jinbin Bai, Zhuoran Zhao, Wenhao Chai, Kaidong Yu, Jianzong Wu, Shuangyong Song, Yunhai Tong, Xiangtai Li, Xuelong Li, Shuicheng Yan
Abstract: Unified generation models aim to handle diverse tasks across modalities -- such as text generation, image generation, and vision-language reasoning -- within a single architecture and decoding paradigm. Autoregressive unified models suffer from slow inference due to sequential decoding, and non-autoregressive unified models suffer from weak generalization due to limited pretrained backbones. We introduce Muddit, a unified discrete diffusion transformer that enables fast and parallel generation across both text and image modalities. Unlike prior unified diffusion models trained from scratch, Muddit integrates strong visual priors from a pretrained text-to-image backbone with a lightweight text decoder, enabling flexible and high-quality multimodal generation under a unified architecture. Empirical results show that Muddit achieves competitive or superior performance compared to significantly larger autoregressive models in both quality and efficiency. The work highlights the potential of purely discrete diffusion, when equipped with strong visual priors, as a scalable and effective backbone for unified generation.
Authors: Carolina Fortuna, Gregor Cerar, Blaz Bertalanic, Andrej Campa, Mihael Mohorcic
Abstract: The transition from traditional power grids to smart grids, significant increase in the use of renewable energy sources, and soaring electricity prices has triggered a digital transformation of the energy infrastructure that enables new, data driven, applications often supported by machine learning models. However, the majority of the developed machine learning models rely on univariate data. To date, a structured study considering the role meta-data and additional measurements resulting in multivariate data is missing. In this paper we propose a taxonomy that identifies and structures various types of data related to energy applications. The taxonomy can be used to guide application specific data model development for training machine learning models. Focusing on a household electricity forecasting application, we validate the effectiveness of the proposed taxonomy in guiding the selection of the features for various types of models. As such, we study of the effect of domain, contextual and behavioral features on the forecasting accuracy of four interpretable machine learning techniques and three openly available datasets. Finally, using a feature importance techniques, we explain individual feature contributions to the forecasting accuracy.
Authors: Armando Bellante, Martin Pl\'avala, Alessandro Luongo
Abstract: This paper proposes a family of permutation-invariant graph embeddings, generalizing the Skew Spectrum of graphs of Kondor & Borgwardt (2008). Grounded in group theory and harmonic analysis, our method introduces a new class of graph invariants that are isomorphism-invariant and capable of embedding richer graph structures - including attributed graphs, multilayer graphs, and hypergraphs - which the Skew Spectrum could not handle. Our generalization further defines a family of functions that enables a trade-off between computational complexity and expressivity. By applying generalization-preserving heuristics to this family, we improve the Skew Spectrum's expressivity at the same computational cost. We formally prove the invariance of our generalization, demonstrate its improved expressiveness through experiments, and discuss its efficient computation.
Authors: Xiangcheng Zhang, Haowei Lin, Haotian Ye, James Zou, Jianzhu Ma, Yitao Liang, Yilun Du
Abstract: Classical search algorithms have long underpinned modern artificial intelligence. In this work, we tackle the challenge of inference-time control in diffusion models -- adapting generated outputs to meet diverse test-time objectives -- using principles from classical search. We propose a general framework that orchestrates local and global search to efficiently navigate the generative space. It employs a theoretically grounded local search via annealed Langevin MCMC and performs compute-efficient global exploration using breadth-first and depth-first tree search. We evaluate our approach on a range of challenging domains, including planning, offline reinforcement learning, and image generation. Across all tasks, we observe significant gains in both performance and efficiency. These results show that classical search provides a principled and practical foundation for inference-time scaling in diffusion models. Project page at diffusion-inference-scaling.github.io.
Authors: Chang Yue, Niraj K. Jha
Abstract: Neural networks (NNs) achieve outstanding performance in many domains; however, their decision processes are often opaque and their inference can be computationally expensive in resource-constrained environments. We recently proposed Differentiable Logic Networks (DLNs) to address these issues for tabular classification based on relaxing discrete logic into a differentiable form, thereby enabling gradient-based learning of networks built from binary logic operations. DLNs offer interpretable reasoning and substantially lower inference cost. We extend the DLN framework to supervised tabular regression. Specifically, we redesign the final output layer to support continuous targets and unify the original two-phase training procedure into a single differentiable stage. We evaluate the resulting model on 15 public regression benchmarks, comparing it with modern neural networks and classical regression baselines. Regression DLNs match or exceed baseline accuracy while preserving interpretability and fast inference. Our results show that DLNs are a viable, cost-effective alternative for regression tasks, especially where model transparency and computational efficiency are important.
Authors: Griffin Dietz Smith, Dianna Yee, Jennifer King Chen, Leah Findlater
Abstract: Identifying mistakes (i.e., miscues) made while reading aloud is commonly approached post-hoc by comparing automatic speech recognition (ASR) transcriptions to the target reading text. However, post-hoc methods perform poorly when ASR inaccurately transcribes verbatim speech. To improve on current methods for reading error annotation, we propose a novel end-to-end architecture that incorporates the target reading text via prompting and is trained for both improved verbatim transcription and direct miscue detection. Our contributions include: first, demonstrating that incorporating reading text through prompting benefits verbatim transcription performance over fine-tuning, and second, showing that it is feasible to augment speech recognition tasks for end-to-end miscue detection. We conducted two case studies -- children's read-aloud and adult atypical speech -- and found that our proposed strategies improve verbatim transcription and miscue detection compared to current state-of-the-art.
Authors: John Halloran
Abstract: The model context protocol (MCP) has been widely adapted as an open standard enabling the seamless integration of generative AI agents. However, recent work has shown the MCP is susceptible to retrieval-based "falsely benign" attacks (FBAs), allowing malicious system access and credential theft, but requiring that users download compromised files directly to their systems. Herein, we show that the threat model of MCP-based attacks is significantly broader than previously thought, i.e., attackers need only post malicious content online to deceive MCP agents into carrying out their attacks on unsuspecting victims' systems. To improve alignment guardrails against such attacks, we introduce a new MCP dataset of FBAs and (truly) benign samples to explore the effectiveness of direct preference optimization (DPO) for the refusal training of large language models (LLMs). While DPO improves model guardrails against such attacks, we show that the efficacy of refusal learning varies drastically depending on the model's original post-training alignment scheme--e.g., GRPO-based LLMs learn to refuse extremely poorly. Thus, to further improve FBA refusals, we introduce Retrieval Augmented Generation for Preference alignment (RAG-Pref), a novel preference alignment strategy based on RAG. We show that RAG-Pref significantly improves the ability of LLMs to refuse FBAs, particularly when combined with DPO alignment, thus drastically improving guardrails against MCP-based attacks.
Authors: Yilin Xie, Shiqiang Zhang, Jixiang Qing, Ruth Misener, Calvin Tsay
Abstract: Graph Bayesian optimization (BO) has shown potential as a powerful and data-efficient tool for neural architecture search (NAS). Most existing graph BO works focus on developing graph surrogates models, i.e., metrics of networks and/or different kernels to quantify the similarity between networks. However, the acquisition optimization, as a discrete optimization task over graph structures, is not well studied due to the complexity of formulating the graph search space and acquisition functions. This paper presents explicit optimization formulations for graph input space including properties such as reachability and shortest paths, which are used later to formulate graph kernels and the acquisition function. We theoretically prove that the proposed encoding is an equivalent representation of the graph space and provide restrictions for the NAS domain with either node or edge labels. Numerical results over several NAS benchmarks show that our method efficiently finds the optimal architecture for most cases, highlighting its efficacy.
Authors: Halil Alperen Gozeten, M. Emrullah Ildiz, Xuechen Zhang, Hrayr Harutyunyan, Ankit Singh Rawat, Samet Oymak
Abstract: Current language models generate chain-of-thought traces by autoregressively sampling tokens from a finite vocabulary. While this discrete sampling has achieved remarkable success, conducting chain-of-thought with continuously-valued tokens (CoT2) offers a richer and more expressive alternative. Our work examines the benefits of CoT2 through logical reasoning tasks that inherently require search capabilities and provide optimization and exploration methods for CoT2. Theoretically, we show that CoT2 allows the model to track multiple traces in parallel and quantify its benefits for inference efficiency. Notably, one layer transformer equipped with CoT2 can provably solve the combinatorial "subset sum problem" given sufficient embedding dimension. These insights lead to a novel and effective supervision strategy where we match the softmax outputs to the empirical token distributions of a set of target traces. Complementing this, we introduce sampling strategies that unlock policy optimization and self-improvement for CoT2. Our first strategy samples and composes $K$ discrete tokens at each decoding step to control the level of parallelism, and reduces to standard CoT when $K=1$. Our second strategy relies on continuous exploration over the probability simplex. Experiments confirm that policy optimization with CoT2 indeed improves the performance of the model beyond its initial discrete or continuous supervision.
Authors: Juncheol Shin, Minsang Seok, Seonggon Kim, Eunhyeok Park
Abstract: Model merging has emerged as a powerful technique for combining task-specific weights, achieving superior performance in multi-target domain adaptation. However, when applied to practical scenarios, such as quantized models, new challenges arise. In practical scenarios, quantization is often applied to target-specific data, but this process restricts the domain of interest and introduces discretization effects, making model merging highly non-trivial. In this study, we analyze the impact of quantization on model merging through the lens of error barriers. Leveraging these insights, we propose a novel post-training quantization, HDRQ - Hessian and distant regularizing quantization - that is designed to consider model merging for multi-target domain adaptation. Our approach ensures that the quantization process incurs minimal deviation from the source pre-trained model while flattening the loss surface to facilitate smooth model merging. To our knowledge, this is the first study on this challenge, and extensive experiments confirm its effectiveness.
Authors: Jiaran Ye, Zijun Yao, Zhidian Huang, Liangming Pan, Jinxin Liu, Yushi Bai, Amy Xin, Liu Weichuan, Xiaoyin Che, Lei Hou, Juanzi Li
Abstract: Recent work suggests that large language models (LLMs) can perform multi-hop reasoning implicitly -- producing correct answers without explicitly verbalizing intermediate steps -- but the underlying mechanisms remain poorly understood. In this paper, we study how such implicit reasoning emerges by training transformers from scratch in a controlled symbolic environment. Our analysis reveals a three-stage developmental trajectory: early memorization, followed by in-distribution generalization, and eventually cross-distribution generalization. We find that training with atomic triples is not necessary but accelerates learning, and that second-hop generalization relies on query-level exposure to specific compositional structures. To interpret these behaviors, we introduce two diagnostic tools: cross-query semantic patching, which identifies semantically reusable intermediate representations, and a cosine-based representational lens, which reveals that successful reasoning correlates with the cosine-base clustering in hidden space. This clustering phenomenon in turn provides a coherent explanation for the behavioral dynamics observed across training, linking representational structure to reasoning capability. These findings provide new insights into the interpretability of implicit multi-hop reasoning in LLMs, helping to clarify how complex reasoning processes unfold internally and offering pathways to enhance the transparency of such models.
Authors: Niklas Freymuth, Tobias W\"urth, Nicolas Schreiber, Balazs Gyenes, Andreas Boltres, Johannes Mitsch, Aleksandar Taranovic, Tai Hoang, Philipp Dahlinger, Philipp Becker, Luise K\"arger, Gerhard Neumann
Abstract: The cost and accuracy of simulating complex physical systems using the Finite Element Method (FEM) scales with the resolution of the underlying mesh. Adaptive meshes improve computational efficiency by refining resolution in critical regions, but typically require task-specific heuristics or cumbersome manual design by a human expert. We propose Adaptive Meshing By Expert Reconstruction (AMBER), a supervised learning approach to mesh adaptation. Starting from a coarse mesh, AMBER iteratively predicts the sizing field, i.e., a function mapping from the geometry to the local element size of the target mesh, and uses this prediction to produce a new intermediate mesh using an out-of-the-box mesh generator. This process is enabled through a hierarchical graph neural network, and relies on data augmentation by automatically projecting expert labels onto AMBER-generated data during training. We evaluate AMBER on 2D and 3D datasets, including classical physics problems, mechanical components, and real-world industrial designs with human expert meshes. AMBER generalizes to unseen geometries and consistently outperforms multiple recent baselines, including ones using Graph and Convolutional Neural Networks, and Reinforcement Learning-based approaches.
Authors: Aya Kayal, Sattar Vakili, Laura Toni, Da-shan Shiu, Alberto Bernacchia
Abstract: Bayesian optimization (BO) with preference-based feedback has recently garnered significant attention due to its emerging applications. We refer to this problem as Bayesian Optimization from Human Feedback (BOHF), which differs from conventional BO by learning the best actions from a reduced feedback model, where only the preference between two actions is revealed to the learner at each time step. The objective is to identify the best action using a limited number of preference queries, typically obtained through costly human feedback. Existing work, which adopts the Bradley-Terry-Luce (BTL) feedback model, provides regret bounds for the performance of several algorithms. In this work, within the same framework we develop tighter performance guarantees. Specifically, we derive regret bounds of $\tilde{\mathcal{O}}(\sqrt{\Gamma(T)T})$, where $\Gamma(T)$ represents the maximum information gain$\unicode{x2014}$a kernel-specific complexity term$\unicode{x2014}$and $T$ is the number of queries. Our results significantly improve upon existing bounds. Notably, for common kernels, we show that the order-optimal sample complexities of conventional BO$\unicode{x2014}$achieved with richer feedback models$\unicode{x2014}$are recovered. In other words, the same number of preferential samples as scalar-valued samples is sufficient to find a nearly optimal solution.
Authors: Bo Zhao, Nima Dehmamy, Robin Walters, Rose Yu
Abstract: Neural network minima are often connected by curves along which train and test loss remain nearly constant, a phenomenon known as mode connectivity. While this property has enabled applications such as model merging and fine-tuning, its theoretical explanation remains unclear. We propose a new approach to exploring the connectedness of minima using parameter space symmetry. By linking the topology of symmetry groups to that of the minima, we derive the number of connected components of the minima of linear networks and show that skip connections reduce this number. We then examine when mode connectivity and linear mode connectivity hold or fail, using parameter symmetries which account for a significant part of the minimum. Finally, we provide explicit expressions for connecting curves in the minima induced by symmetry. Using the curvature of these curves, we derive conditions under which linear mode connectivity approximately holds. Our findings highlight the role of continuous symmetries in understanding the neural network loss landscape.
Authors: Zixuan Wang, Eshaan Nichani, Alberto Bietti, Alex Damian, Daniel Hsu, Jason D. Lee, Denny Wu
Abstract: Transformer-based language models have demonstrated impressive capabilities across a range of complex reasoning tasks. Prior theoretical work exploring the expressive power of transformers has shown that they can efficiently perform multi-step reasoning tasks involving parallelizable computations. However, the learnability of such constructions, particularly the conditions on the data distribution that enable efficient learning via gradient-based optimization, remains an open question. Towards answering this question, in this work we study the learnability of the $k$-fold composition task, which requires computing an interleaved composition of $k$ input permutations and $k$ hidden permutations, and can be expressed by a transformer with $O(\log k)$ layers. On the negative front, we prove a Statistical Query (SQ) lower bound showing that any SQ learner that makes only polynomially-many queries to an SQ oracle for the $k$-fold composition task distribution must have sample size exponential in $k$, thus establishing a statistical-computational gap. On the other hand, we show that this function class can be efficiently learned, with runtime and sample complexity polynomial in $k$, by gradient descent on an $O(\log k)$-depth transformer via two different curriculum learning strategies: one in which data consists of $k'$-fold composition functions with $k' \le k$ presented in increasing difficulty, and another in which all such data is presented simultaneously. Our work sheds light on the necessity and sufficiency of having both easy and hard examples in the data distribution for transformers to learn complex compositional tasks.
Authors: Hiroshi Kera, Nico Pelleriti, Yuki Ishihara, Max Zimmer, Sebastian Pokutta
Abstract: Solving systems of polynomial equations, particularly those with finitely many solutions, is a crucial challenge across many scientific fields. Traditional methods like Gr\"obner and Border bases are fundamental but suffer from high computational costs, which have motivated recent Deep Learning approaches to improve efficiency, albeit at the expense of output correctness. In this work, we introduce the Oracle Border Basis Algorithm, the first Deep Learning approach that accelerates Border basis computation while maintaining output guarantees. To this end, we design and train a Transformer-based oracle that identifies and eliminates computationally expensive reduction steps, which we find to dominate the algorithm's runtime. By selectively invoking this oracle during critical phases of computation, we achieve substantial speedup factors of up to 3.5x compared to the base algorithm, without compromising the correctness of results. To generate the training data, we develop a sampling method and provide the first sampling theorem for border bases. We construct a tokenization and embedding scheme tailored to monomial-centered algebraic computations, resulting in a compact and expressive input representation, which reduces the number of tokens to encode an $n$-variate polynomial by a factor of $O(n)$. Our learning approach is data efficient, stable, and a practical enhancement to traditional computer algebra algorithms and symbolic computation.
Authors: Oleksii Furman, Ulvi Movsum-zada, Patryk Marszalek, Maciej Zi\k{e}ba, Marek \'Smieja
Abstract: Counterfactual explanations play a pivotal role in explainable artificial intelligence (XAI) by offering intuitive, human-understandable alternatives that elucidate machine learning model decisions. Despite their significance, existing methods for generating counterfactuals often require constant access to the predictive model, involve computationally intensive optimization for each instance and lack the flexibility to adapt to new user-defined constraints without retraining. In this paper, we propose DiCoFlex, a novel model-agnostic, conditional generative framework that produces multiple diverse counterfactuals in a single forward pass. Leveraging conditional normalizing flows trained solely on labeled data, DiCoFlex addresses key limitations by enabling real-time user-driven customization of constraints such as sparsity and actionability at inference time. Extensive experiments on standard benchmark datasets show that DiCoFlex outperforms existing methods in terms of validity, diversity, proximity, and constraint adherence, making it a practical and scalable solution for counterfactual generation in sensitive decision-making domains.
Authors: Nathan Lichtl\'e, Alexi Canesse, Zhe Fu, Hossein Nick Zinat Matin, Maria Laura Delle Monache, Alexandre M. Bayen
Abstract: We introduce (U)NFV, a modular neural network architecture that generalizes classical finite volume (FV) methods for solving hyperbolic conservation laws. Hyperbolic partial differential equations (PDEs) are challenging to solve, particularly conservation laws whose physically relevant solutions contain shocks and discontinuities. FV methods are widely used for their mathematical properties: convergence to entropy solutions, flow conservation, or total variation diminishing, but often lack accuracy and flexibility in complex settings. Neural Finite Volume addresses these limitations by learning update rules over extended spatial and temporal stencils while preserving conservation structure. It supports both supervised training on solution data (NFV) and unsupervised training via weak-form residual loss (UNFV). Applied to first-order conservation laws, (U)NFV achieves up to 10x lower error than Godunov's method, outperforms ENO/WENO, and rivals discontinuous Galerkin solvers with far less complexity. On traffic modeling problems, both from PDEs and from experimental highway data, (U)NFV captures nonlinear wave dynamics with significantly higher fidelity and scalability than traditional FV approaches.
Authors: Danny Driess, Jost Tobias Springenberg, Brian Ichter, Lili Yu, Adrian Li-Bell, Karl Pertsch, Allen Z. Ren, Homer Walke, Quan Vuong, Lucy Xiaoyang Shi, Sergey Levine
Abstract: Vision-language-action (VLA) models provide a powerful approach to training control policies for physical systems, such as robots, by combining end-to-end learning with transfer of semantic knowledge from web-scale vision-language model (VLM) training. However, the constraints of real-time control are often at odds with the design of VLMs: the most powerful VLMs have tens or hundreds of billions of parameters, presenting an obstacle to real-time inference, and operate on discrete tokens rather than the continuous-valued outputs that are required for controlling robots. To address this challenge, recent VLA models have used specialized modules for efficient continuous control, such as action experts or continuous output heads, which typically require adding new untrained parameters to the pretrained VLM backbone. While these modules improve real-time and control capabilities, it remains an open question whether they preserve or degrade the semantic knowledge contained in the pretrained VLM, and what effect they have on the VLA training dynamics. In this paper, we study this question in the context of VLAs that include a continuous diffusion or flow matching action expert, showing that naively including such experts significantly harms both training speed and knowledge transfer. We provide an extensive analysis of various design choices, their impact on performance and knowledge transfer, and propose a technique for insulating the VLM backbone during VLA training that mitigates this issue. Videos are available at https://pi.website/research/knowledge_insulation.
Authors: Andreas Auer, Patrick Podest, Daniel Klotz, Sebastian B\"ock, G\"unter Klambauer, Sepp Hochreiter
Abstract: In-context learning, the ability of large language models to perform tasks using only examples provided in the prompt, has recently been adapted for time series forecasting. This paradigm enables zero-shot prediction, where past values serve as context for forecasting future values, making powerful forecasting tools accessible to non-experts and increasing the performance when training data are scarce. Most existing zero-shot forecasting approaches rely on transformer architectures, which, despite their success in language, often fall short of expectations in time series forecasting, where recurrent models like LSTMs frequently have the edge. Conversely, while LSTMs are well-suited for time series modeling due to their state-tracking capabilities, they lack strong in-context learning abilities. We introduce TiRex that closes this gap by leveraging xLSTM, an enhanced LSTM with competitive in-context learning skills. Unlike transformers, state-space models, or parallelizable RNNs such as RWKV, TiRex retains state-tracking, a critical property for long-horizon forecasting. To further facilitate its state-tracking ability, we propose a training-time masking strategy called CPM. TiRex sets a new state of the art in zero-shot time series forecasting on the HuggingFace benchmarks GiftEval and Chronos-ZS, outperforming significantly larger models including TabPFN-TS (Prior Labs), Chronos Bolt (Amazon), TimesFM (Google), and Moirai (Salesforce) across both short- and long-term forecasts.
Authors: Arun Verma, Indrajit Saha, Makoto Yokoo, Bryan Kian Hsiang Low
Abstract: This paper considers a contextual bandit problem involving multiple agents, where a learner sequentially observes the contexts and the agent's reported arms, and then selects the arm that maximizes the system's overall reward. Existing work in contextual bandits assumes that agents truthfully report their arms, which is unrealistic in many real-life applications. For instance, consider an online platform with multiple sellers; some sellers may misrepresent product quality to gain an advantage, such as having the platform preferentially recommend their products to online users. To address this challenge, we propose an algorithm, COBRA, for contextual bandit problems involving strategic agents that disincentivize their strategic behavior without using any monetary incentives, while having incentive compatibility and a sub-linear regret guarantee. Our experimental results also validate the different performance aspects of our proposed algorithm.
Authors: Sean Current, Ziqi Chen, Daniel Adu-Ampratwum, Xia Ning, Srinivasan Parthasarathy
Abstract: Methods for automatic chemical retrosynthesis have found recent success through the application of models traditionally built for natural language processing, primarily through transformer neural networks. These models have demonstrated significant ability to translate between the SMILES encodings of chemical products and reactants, but are constrained as a result of their autoregressive nature. We propose DiffER, an alternative template-free method for retrosynthesis prediction in the form of categorical diffusion, which allows the entire output SMILES sequence to be predicted in unison. We construct an ensemble of diffusion models which achieves state-of-the-art performance for top-1 accuracy and competitive performance for top-3, top-5, and top-10 accuracy among template-free methods. We prove that DiffER is a strong baseline for a new class of template-free model, capable of learning a variety of synthetic techniques used in laboratory settings and outperforming a variety of other template-free methods on top-k accuracy metrics. By constructing an ensemble of categorical diffusion models with a novel length prediction component with variance, our method is able to approximately sample from the posterior distribution of reactants, producing results with strong metrics of confidence and likelihood. Furthermore, our analyses demonstrate that accurate prediction of the SMILES sequence length is key to further boosting the performance of categorical diffusion models.
Authors: Minrui Luo, Fuhang Kuang, Yu Wang, Zirui Liu, Tianxing He
Abstract: Parameter-Efficient Fine-Tuning (PEFT) methods, particularly Low-Rank Adaptation (LoRA), are indispensable for efficiently customizing Large Language Models (LLMs). However, vanilla LoRA suffers from slow convergence speed and knowledge forgetting problems. Recent studies have leveraged the power of designed LoRA initialization, to enhance the fine-tuning efficiency, or to preserve knowledge in the pre-trained LLM. However, none of these works can address the two cases at the same time. To this end, we introduce Subspace-Constrained LoRA (SC-LoRA), a novel LoRA initialization framework engineered to navigate the trade-off between efficient fine-tuning and knowledge preservation. We achieve this by constraining the output of trainable LoRA adapters in a low-rank subspace, where the context information of fine-tuning data is most preserved while the context information of preserved knowledge is least retained, in a balanced way. Such constraint enables the trainable weights to primarily focus on the main features of fine-tuning data while avoiding damaging the preserved knowledge features. We provide theoretical analysis on our method, and conduct extensive experiments including safety preservation and world knowledge preservation, on various downstream tasks. In our experiments, SC-LoRA succeeds in delivering superior fine-tuning performance while markedly diminishing knowledge forgetting, surpassing contemporary LoRA initialization methods.
Authors: Benjamin Th\'erien, Xiaolong Huang, Irina Rish, Eugene Belilovsky
Abstract: DiLoCo is a powerful framework for training large language models (LLMs) under networking constraints with advantages for increasing parallelism and accelerator utilization in data center settings. Despite significantly reducing communication frequency, however, DiLoCo's communication steps still involve all-reducing a complete copy of the model's parameters. While existing works have explored ways to reduce communication in DiLoCo, the role of error feedback accumulators and the effect of the inner-optimizer on compressibility remain under-explored. In this work, we investigate the effectiveness of standard compression methods including Top-k sparsification and quantization for reducing the communication overhead of DiLoCo when paired with two local optimizers (AdamW and Muon). Our experiments pre-training decoder-only transformer language models (LMs) reveal that leveraging Muon as the inner optimizer for DiLoCo along with an error-feedback accumulator allows to aggressively compress the communicated delta to 2-bits with next to no performance degradation. Crucially, MuLoCo (Muon inner optimizer DiLoCo) significantly outperforms DiLoCo while communicating 8X less and having identical memory complexity.
Authors: Shreeram Suresh Chandra, Lucas Goncalves, Junchen Lu, Carlos Busso, Berrak Sisman
Abstract: Current emotion-based contrastive language-audio pretraining (CLAP) methods typically learn by na\"ively aligning audio samples with corresponding text prompts. Consequently, this approach fails to capture the ordinal nature of emotions, hindering inter-emotion understanding and often resulting in a wide modality gap between the audio and text embeddings due to insufficient alignment. To handle these drawbacks, we introduce EmotionRankCLAP, a supervised contrastive learning approach that uses dimensional attributes of emotional speech and natural language prompts to jointly capture fine-grained emotion variations and improve cross-modal alignment. Our approach utilizes a Rank-N-Contrast objective to learn ordered relationships by contrasting samples based on their rankings in the valence-arousal space. EmotionRankCLAP outperforms existing emotion-CLAP methods in modeling emotion ordinality across modalities, measured via a cross-modal retrieval task.
Authors: Paul G\"olz, Nika Haghtalab, Kunhe Yang
Abstract: After pre-training, large language models are aligned with human preferences based on pairwise comparisons. State-of-the-art alignment methods (such as PPO-based RLHF and DPO) are built on the assumption of aligning with a single preference model, despite being deployed in settings where users have diverse preferences. As a result, it is not even clear that these alignment methods produce models that satisfy users on average -- a minimal requirement for pluralistic alignment. Drawing on social choice theory and modeling users' comparisons through individual Bradley-Terry (BT) models, we introduce an alignment method's distortion: the worst-case ratio between the optimal achievable average utility, and the average utility of the learned policy. The notion of distortion helps draw sharp distinctions between alignment methods: Nash Learning from Human Feedback achieves the minimax optimal distortion of $(\frac{1}{2} + o(1)) \cdot \beta$ (for the BT temperature $\beta$), robustly across utility distributions, distributions of comparison pairs, and permissible KL divergences from the reference policy. RLHF and DPO, by contrast, suffer $\geq (1 - o(1)) \cdot \beta$ distortion already without a KL constraint, and $e^{\Omega(\beta)}$ or even unbounded distortion in the full setting, depending on how comparison pairs are sampled.
Authors: Declan Kutscher, David M. Chan, Yutong Bai, Trevor Darrell, Ritwik Gupta
Abstract: Sequence models such as transformers require inputs to be represented as one-dimensional sequences. In vision, this typically involves flattening images using a fixed row-major (raster-scan) order. While full self-attention is permutation-equivariant, modern long-sequence transformers increasingly rely on architectural approximations that break this invariance and introduce sensitivity to patch ordering. We show that patch order significantly affects model performance in such settings, with simple alternatives like column-major or Hilbert curves yielding notable accuracy shifts. Motivated by this, we propose REOrder, a two-stage framework for discovering task-optimal patch orderings. First, we derive an information-theoretic prior by evaluating the compressibility of various patch sequences. Then, we learn a policy over permutations by optimizing a Plackett-Luce policy using REINFORCE. This approach enables efficient learning in a combinatorial permutation space. REOrder improves top-1 accuracy over row-major ordering on ImageNet-1K by up to 3.01% and Functional Map of the World by 13.35%.
Authors: Amber Yijia Zheng, Cedar Site Bai, Brian Bullins, Raymond A. Yeh
Abstract: Model immunization aims to pre-train models that are difficult to fine-tune on harmful tasks while retaining their utility on other non-harmful tasks. Though prior work has shown empirical evidence for immunizing text-to-image models, the key understanding of when immunization is possible and a precise definition of an immunized model remain unclear. In this work, we propose a framework, based on the condition number of a Hessian matrix, to analyze model immunization for linear models. Building on this framework, we design an algorithm with regularization terms to control the resulting condition numbers after pre-training. Empirical results on linear models and non-linear deep-nets demonstrate the effectiveness of the proposed algorithm on model immunization. The code is available at https://github.com/amberyzheng/model-immunization-cond-num.
URLs: https://github.com/amberyzheng/model-immunization-cond-num.
Authors: Yunjae Won, Hyunji Lee, Hyeonbin Hwang, Minjoon Seo
Abstract: Direct Preference Optimization (DPO) has become a standard technique for aligning language models with human preferences in a supervised manner. Despite its empirical success, the theoretical justification behind its log-ratio reward parameterization remains incomplete. In this work, we address this gap by utilizing the Differential Information Distribution (DID): a distribution over token sequences that captures the information gained during policy updates. First, we show that when preference labels encode the differential information required to transform a reference policy into a target policy, the log-ratio reward in DPO emerges as the uniquely optimal form for learning the target policy via preference optimization. This result naturally yields a closed-form expression for the optimal sampling distribution over rejected responses. Second, we find that the condition for preferences to encode differential information is fundamentally linked to an implicit assumption regarding log-margin ordered policies-an inductive bias widely used in preference optimization yet previously unrecognized. Finally, by analyzing the entropy of the DID, we characterize how learning low-entropy differential information reinforces the policy distribution, while high-entropy differential information induces a smoothing effect, which explains the log-likelihood displacement phenomenon. We validate our theoretical findings in synthetic experiments and extend them to real-world instruction-following datasets. Our results suggest that learning high-entropy differential information is crucial for general instruction-following, while learning low-entropy differential information benefits knowledge-intensive question answering. Overall, our work presents a unifying perspective on the DPO objective, the structure of preference data, and resulting policy behaviors through the lens of differential information.
Authors: Arghya Mallick, Georgios Pantazis, Mohammad Khosravi, Peyman Mohajerin Esfahani, Sergio Grammatico
Abstract: We propose a data-driven, user-centric vehicle-to-grid (V2G) methodology based on multi-objective optimization to balance battery degradation and V2G revenue according to EV user preference. Given the lack of accurate and generalizable battery degradation models, we leverage input convex neural networks (ICNNs) to develop a data-driven degradation model trained on extensive experimental datasets. This approach enables our model to capture nonconvex dependencies on battery temperature and time while maintaining convexity with respect to the charging rate. Such a partial convexity property ensures that the second stage of our methodology remains computationally efficient. In the second stage, we integrate our data-driven degradation model into a multi-objective optimization framework to generate an optimal smart charging profile for each EV. This profile effectively balances the trade-off between financial benefits from V2G participation and battery degradation, controlled by a hyperparameter reflecting the user prioritization of battery health. Numerical simulations show the high accuracy of the ICNN model in predicting battery degradation for unseen data. Finally, we present a trade-off curve illustrating financial benefits from V2G versus losses from battery health degradation based on user preferences and showcase smart charging strategies under realistic scenarios.
Authors: Jin Han, Xin-Zheng Lu, Jia-Rui Lin
Abstract: Building Information Modeling (BIM) has revolutionized the construction industry by providing a comprehensive digital representation of building structures throughout their lifecycle. However, existing research lacks effective methods for capturing the complex spatial and topological relationships between components in BIM models, which are essential for understanding design patterns and enhancing decision-making. This study proposes a unified network-based representation method that integrates the "semantic-spatial-topological" multi-dimensional design features of BIM models. By extending the IFC (Industry Foundation Classes) standard, we introduce local spatial relationships and topological connections between components to enrich the network structure. This representation method enables a more detailed understanding of component interactions, dependencies, and implicit design patterns, effectively capturing the semantic, topological, and spatial relationships in BIM, and holds significant potential for the representation and learning of design patterns.
Authors: Pawan Neupane, Jian Liu, Jianlin Cheng
Abstract: Predicting protein complex structures is essential for protein function analysis, protein design, and drug discovery. While AI methods like AlphaFold can predict accurate structural models for many protein complexes, reliably estimating the quality of these predicted models (estimation of model accuracy, or EMA) for model ranking and selection remains a major challenge. A key barrier to developing effective machine learning-based EMA methods is the lack of large, diverse, and well-annotated datasets for training and evaluation. To address this gap, we introduce PSBench, a benchmark suite comprising four large-scale, labeled datasets generated during the 15th and 16th community-wide Critical Assessment of Protein Structure Prediction (CASP15 and CASP16). PSBench includes over one million structural models covering a wide range of protein sequence lengths, complex stoichiometries, functional classes, and modeling difficulties. Each model is annotated with multiple complementary quality scores at the global, local, and interface levels. PSBench also provides multiple evaluation metrics and baseline EMA methods to facilitate rigorous comparisons. To demonstrate PSBench's utility, we trained and evaluated GATE, a graph transformer-based EMA method, on the CASP15 data. GATE was blindly tested in CASP16 (2024), where it ranked among the top-performing EMA methods. These results highlight PSBench as a valuable resource for advancing EMA research in protein complex modeling. PSBench is publicly available at: https://github.com/BioinfoMachineLearning/PSBench.
Authors: Jiahao Yang, Ran Fang, Ming Zhang, Jun Zhou
Abstract: In high-frequency trading (HFT), leveraging limit order books (LOB) to model stock price movements is crucial for achieving profitable outcomes. However, this task is challenging due to the high-dimensional and volatile nature of the original data. Even recent deep learning models often struggle to capture price movement patterns effectively, particularly without well-designed features. We observed that raw LOB data exhibits inherent symmetry between the ask and bid sides, and the bid-ask differences demonstrate greater stability and lower complexity compared to the original data. Building on this insight, we propose a novel approach in which leverages the Siamese architecture to enhance the performance of existing deep learning models. The core idea involves processing the ask and bid sides separately using the same module with shared parameters. We applied our Siamese-based methods to several widely used strong baselines and validated their effectiveness using data from 14 military industry stocks in the Chinese A-share market. Furthermore, we integrated multi-head attention (MHA) mechanisms with the Long Short-Term Memory (LSTM) module to investigate its role in modeling stock price movements. Our experiments used raw data and widely used Order Flow Imbalance (OFI) features as input with some strong baseline models. The results show that our method improves the performance of strong baselines in over 75$% of cases, excluding the Multi-Layer Perception (MLP) baseline, which performed poorly and is not considered practical. Furthermore, we found that Multi-Head Attention can enhance model performance, particularly over shorter forecasting horizons.
Authors: Yufeng Wang, Yiguang Bai, Tianqing Zhu, Ismail Ben Ayed, Jing Yuan
Abstract: Community partitioning is crucial in network analysis, with modularity optimization being the prevailing technique. However, traditional modularity-based methods often overlook fairness, a critical aspect in real-world applications. To address this, we introduce protected group networks and propose a novel fairness-modularity metric. This metric extends traditional modularity by explicitly incorporating fairness, and we prove that minimizing it yields naturally fair partitions for protected groups while maintaining theoretical soundness. We develop a general optimization framework for fairness partitioning and design the efficient Fair Fast Newman (FairFN) algorithm, enhancing the Fast Newman (FN) method to optimize both modularity and fairness. Experiments show FairFN achieves significantly improved fairness and high-quality partitions compared to state-of-the-art methods, especially on unbalanced datasets.
Authors: Palur Venkata Raghuvamsi, Siyuan Brandon Loh, Prasanta Bhattacharya, Joses Ho, Raphael Lee Tze Chuen, Alvin X. Han, Sebastian Maurer-Stroh
Abstract: The COVID-19 pandemic response relied heavily on statistical and machine learning models to predict key outcomes such as case prevalence and fatality rates. These predictions were instrumental in enabling timely public health interventions that helped break transmission cycles. While most existing models are grounded in traditional epidemiological data, the potential of alternative datasets, such as those derived from genomic information and human behavior, remains underexplored. In the current study, we investigated the usefulness of diverse modalities of feature sets in predicting case surges. Our results highlight the relative effectiveness of biological (e.g., mutations), public health (e.g., case counts, policy interventions) and human behavioral features (e.g., mobility and social media conversations) in predicting country-level case surges. Importantly, we uncover considerable heterogeneity in predictive performance across countries and feature modalities, suggesting that surge prediction models may need to be tailored to specific national contexts and pandemic phases. Overall, our work highlights the value of integrating alternative data sources into existing disease surveillance frameworks to enhance the prediction of pandemic dynamics.
Authors: Sitan Chen, Weiyuan Gong, Jonas Haferkamp, Yihui Quek
Abstract: In a variety of physically relevant settings for learning from quantum data, designing protocols that can computationally efficiently extract information remains largely an art, and there are important cases where we believe this to be impossible, that is, where there is an information-computation gap. While there is a large array of tools in the classical literature for giving evidence for average-case hardness of statistical inference problems, the corresponding tools in the quantum literature are far more limited. One such framework in the classical literature, the low-degree method, makes predictions about hardness of inference problems based on the failure of estimators given by low-degree polynomials. In this work, we extend this framework to the quantum setting. We establish a general connection between state designs and low-degree hardness. We use this to obtain the first information-computation gaps for learning Gibbs states of random, sparse, non-local Hamiltonians. We also use it to prove hardness for learning random shallow quantum circuit states in a challenging model where states can be measured in adaptively chosen bases. To our knowledge, the ability to model adaptivity within the low-degree framework was open even in classical settings. In addition, we also obtain a low-degree hardness result for quantum error mitigation against strategies with single-qubit measurements. We define a new quantum generalization of the planted biclique problem and identify the threshold at which this problem becomes computationally hard for protocols that perform local measurements. Interestingly, the complexity landscape for this problem shifts when going from local measurements to more entangled single-copy measurements. We show average-case hardness for the "standard" variant of Learning Stabilizers with Noise and for agnostically learning product states.
Authors: Jose Guadalupe Hernandez, Attri Ghosh, Philip J. Freda, Yufei Meng, Nicholas Matsumoto, Jason H. Moore
Abstract: We present the Star-Based Automated Single-locus and Epistasis analysis tool - Genetic Programming (StarBASE-GP), an automated framework for discovering meaningful genetic variants associated with phenotypic variation in large-scale genomic datasets. StarBASE-GP uses a genetic programming-based multi-objective optimization strategy to evolve machine learning pipelines that simultaneously maximize explanatory power (r2) and minimize pipeline complexity. Biological domain knowledge is integrated at multiple stages, including the use of nine inheritance encoding strategies to model deviations from additivity, a custom linkage disequilibrium pruning node that minimizes redundancy among features, and a dynamic variant recommendation system that prioritizes informative candidates for pipeline inclusion. We evaluate StarBASE-GP on a cohort of Rattus norvegicus (brown rat) to identify variants associated with body mass index, benchmarking its performance against a random baseline and a biologically naive version of the tool. StarBASE-GP consistently evolves Pareto fronts with superior performance, yielding higher accuracy in identifying both ground truth and novel quantitative trait loci, highlighting relevant targets for future validation. By incorporating evolutionary search and relevant biological theory into a flexible automated machine learning framework, StarBASE-GP demonstrates robust potential for advancing variant discovery in complex traits.
Authors: Tamas Spisak, Karl Friston
Abstract: Attractor dynamics are a hallmark of many complex systems, including the brain. Understanding how such self-organizing dynamics emerge from first principles is crucial for advancing our understanding of neuronal computations and the design of artificial intelligence systems. Here we formalize how attractor networks emerge from the free energy principle applied to a universal partitioning of random dynamical systems. Our approach obviates the need for explicitly imposed learning and inference rules and identifies emergent, but efficient and biologically plausible inference and learning dynamics for such self-organizing systems. These result in a collective, multi-level Bayesian active inference process. Attractors on the free energy landscape encode prior beliefs; inference integrates sensory data into posterior beliefs; and learning fine-tunes couplings to minimize long-term surprise. Analytically and via simulations, we establish that the proposed networks favor approximately orthogonalized attractor representations, a consequence of simultaneously optimizing predictive accuracy and model complexity. These attractors efficiently span the input subspace, enhancing generalization and the mutual information between hidden causes and observable effects. Furthermore, while random data presentation leads to symmetric and sparse couplings, sequential data fosters asymmetric couplings and non-equilibrium steady-state dynamics, offering a natural extension to conventional Boltzmann Machines. Our findings offer a unifying theory of self-organizing attractor networks, providing novel insights for AI and neuroscience.
Authors: Tian Qin, Core Francisco Park, Mujin Kwun, Aaron Walsman, Eran Malach, Nikhil Anand, Hidenori Tanaka, David Alvarez-Melis
Abstract: Mathematical reasoning tasks have become prominent benchmarks for assessing the reasoning capabilities of LLMs, especially with reinforcement learning (RL) methods such as GRPO showing significant performance gains. However, accuracy metrics alone do not support fine-grained assessment of capabilities and fail to reveal which problem-solving skills have been internalized. To better understand these capabilities, we propose to decompose problem solving into fundamental capabilities: Plan (mapping questions to sequences of steps), Execute (correctly performing solution steps), and Verify (identifying the correctness of a solution). Empirically, we find that GRPO mainly enhances the execution skill-improving execution robustness on problems the model already knows how to solve-a phenomenon we call temperature distillation. More importantly, we show that RL-trained models struggle with fundamentally new problems, hitting a 'coverage wall' due to insufficient planning skills. To explore RL's impact more deeply, we construct a minimal, synthetic solution-tree navigation task as an analogy for mathematical problem-solving. This controlled setup replicates our empirical findings, confirming RL primarily boosts execution robustness. Importantly, in this setting, we identify conditions under which RL can potentially overcome the coverage wall through improved exploration and generalization to new solution paths. Our findings provide insights into the role of RL in enhancing LLM reasoning, expose key limitations, and suggest a path toward overcoming these barriers. Code is available at https://github.com/cfpark00/RL-Wall.
Authors: Razvan-Andrei Lascu, Mateusz B. Majka
Abstract: We study the problem of minimizing non-convex functionals on the space of probability measures, regularized by the relative entropy (KL divergence) with respect to a fixed reference measure, as well as the corresponding problem of solving entropy-regularized non-convex-non-concave min-max problems. We utilize the Best Response flow (also known in the literature as the fictitious play flow) and study how its convergence is influenced by the relation between the degree of non-convexity of the functional under consideration, the regularization parameter and the tail behaviour of the reference measure. In particular, we demonstrate how to choose the regularizer, given the non-convex functional, so that the Best Response operator becomes a contraction with respect to the $L^1$-Wasserstein distance, which then ensures the existence of its unique fixed point, which is then shown to be the unique global minimizer for our optimization problem. This extends recent results where the Best Response flow was applied to solve convex optimization problems regularized by the relative entropy with respect to arbitrary reference measures, and with arbitrary values of the regularization parameter. Our results explain precisely how the assumption of convexity can be relaxed, at the expense of making a specific choice of the regularizer. Additionally, we demonstrate how these results can be applied in reinforcement learning in the context of policy optimization for Markov Decision Processes and Markov games with softmax parametrized policies in the mean-field regime.
Authors: Marco Colussi, Dragan Ahmetovic, Sergio Mascetti
Abstract: This paper presents MIAS-SAM, a novel approach for the segmentation of anomalous regions in medical images. MIAS-SAM uses a patch-based memory bank to store relevant image features, which are extracted from normal data using the SAM encoder. At inference time, the embedding patches extracted from the SAM encoder are compared with those in the memory bank to obtain the anomaly map. Finally, MIAS-SAM computes the center of gravity of the anomaly map to prompt the SAM decoder, obtaining an accurate segmentation from the previously extracted features. Differently from prior works, MIAS-SAM does not require to define a threshold value to obtain the segmentation from the anomaly map. Experimental results conducted on three publicly available datasets, each with a different imaging modality (Brain MRI, Liver CT, and Retina OCT) show accurate anomaly segmentation capabilities measured using DICE score. The code is available at: https://github.com/warpcut/MIAS-SAM
Authors: Antonio Ocello, Daniil Tiapkin, Lorenzo Mancini, Mathieu Lauri\`ere, Eric Moulines
Abstract: We introduce Mean-Field Trust Region Policy Optimization (MF-TRPO), a novel algorithm designed to compute approximate Nash equilibria for ergodic Mean-Field Games (MFG) in finite state-action spaces. Building on the well-established performance of TRPO in the reinforcement learning (RL) setting, we extend its methodology to the MFG framework, leveraging its stability and robustness in policy optimization. Under standard assumptions in the MFG literature, we provide a rigorous analysis of MF-TRPO, establishing theoretical guarantees on its convergence. Our results cover both the exact formulation of the algorithm and its sample-based counterpart, where we derive high-probability guarantees and finite sample complexity. This work advances MFG optimization by bridging RL techniques with mean-field decision-making, offering a theoretically grounded approach to solving complex multi-agent problems.
Authors: Charles E. Thornton, Jamie Sloop, Samuel Brown, Aaron Orndorff, William C. Headley, Stephen Young
Abstract: We investigate the end-to-end altitude estimation performance of a convolutional autoencoder-based interference mitigation approach for frequency-modulated continuous-wave (FMCW) radar altimeters. Specifically, we show that a Temporal Convolutional Network (TCN) autoencoder effectively exploits temporal correlations in the received signal, providing superior interference suppression compared to a Least Mean Squares (LMS) adaptive filter. Unlike existing approaches, the present method operates directly on the received FMCW signal. Additionally, we identify key challenges in applying deep learning to wideband FMCW interference mitigation and outline directions for future research to enhance real-time feasibility and generalization to arbitrary interference conditions.
Authors: Siddharth Ancha, Sunshine Jiang, Travis Manderson, Laura Brandt, Yilun Du, Philip R. Osteen, Nicholas Roy
Abstract: In order to navigate safely and reliably in off-road and unstructured environments, robots must detect anomalies that are out-of-distribution (OOD) with respect to the training data. We present an analysis-by-synthesis approach for pixel-wise anomaly detection without making any assumptions about the nature of OOD data. Given an input image, we use a generative diffusion model to synthesize an edited image that removes anomalies while keeping the remaining image unchanged. Then, we formulate anomaly detection as analyzing which image segments were modified by the diffusion model. We propose a novel inference approach for guided diffusion by analyzing the ideal guidance gradient and deriving a principled approximation that bootstraps the diffusion model to predict guidance gradients. Our editing technique is purely test-time that can be integrated into existing workflows without the need for retraining or fine-tuning. Finally, we use a combination of vision-language foundation models to compare pixels in a learned feature space and detect semantically meaningful edits, enabling accurate anomaly detection for off-road navigation. Project website: https://siddancha.github.io/anomalies-by-diffusion-synthesis/
URLs: https://siddancha.github.io/anomalies-by-diffusion-synthesis/
Authors: John C. Duchi
Abstract: The basic question of delineating those statistical problems that are solvable without making any assumptions on the underlying data distribution has long animated statistics and learning theory. This paper characterizes when a (univariate) convex M-estimation or stochastic optimization problem is solvable in such an assumption-free setting, providing a precise dividing line between solvable and unsolvable problems. The conditions we identify show, perhaps surprisingly, that Lipschitz continuity of the loss being minimized is not necessary for distribution free minimization, and they are also distinct from classical characterizations of learnability in machine learning.
Authors: Ba-Hien Tran, Van Minh Nguyen
Abstract: Weight binarization has emerged as a promising strategy to drastically reduce the complexity of large language models (LLMs). It is mainly classified into two approaches: post-training binarization and finetuning with training-aware binarization methods. The first approach, while having low complexity, leads to significant loss of information from the original LLMs, resulting in poor performance. The second approach, on the other hand, relies heavily on full-precision latent weights for gradient approximation of binary weights, which not only remains suboptimal but also introduces substantial complexity. In this paper, we introduce a novel framework that effectively transforms LLMs into multi-kernel Boolean parameters, for the first time, finetunes them directly in the Boolean domain, eliminating the need for expensive latent weights. This significantly reduces complexity during both finetuning and inference. Through extensive and insightful experiments across a wide range of LLMs, we demonstrate that our method outperforms recent ultra low-bit quantization and binarization methods.
Authors: Zhangyi Hu, Jiemin Wu, Hua Xu, Mingqian Liao, Ninghui Feng, Bo Gao, Songning Lai, Yutao Yue
Abstract: Irregular Multivariate Time Series (IMTS) forecasting is challenging due to the unaligned nature of multi-channel signals and the prevalence of extensive missing data. Existing methods struggle to capture reliable temporal patterns from such data due to significant missing values. While pre-trained foundation models show potential for addressing these challenges, they are typically designed for Regularly Sampled Time Series (RTS). Motivated by the visual Mask AutoEncoder's (MAE) powerful capability for modeling sparse multi-channel information and its success in RTS forecasting, we propose VIMTS, a framework adapting Visual MAE for IMTS forecasting. To mitigate the effect of missing values, VIMTS first processes IMTS along the timeline into feature patches at equal intervals. These patches are then complemented using learned cross-channel dependencies. Then it leverages visual MAE's capability in handling sparse multichannel data for patch reconstruction, followed by a coarse-to-fine technique to generate precise predictions from focused contexts. In addition, we integrate self-supervised learning for improved IMTS modeling by adapting the visual MAE to IMTS data. Extensive experiments demonstrate VIMTS's superior performance and few-shot capability, advancing the application of visual foundation models in more general time series tasks. Our code is available at https://github.com/WHU-HZY/VIMTS.
Authors: Arthur S. Bianchessi, Rodrigo C. Barros, Lucas S. Kupssinsk\"u
Abstract: Transformer-based language models rely on positional encoding (PE) to handle token order and support context length extrapolation. However, existing PE methods lack theoretical clarity and rely on limited evaluation metrics to substantiate their extrapolation claims. We propose the Bayesian Attention Mechanism (BAM), a theoretical framework that formulates positional encoding as a prior within a probabilistic model. BAM unifies existing methods (e.g., NoPE and ALiBi) and motivates a new Generalized Gaussian positional prior that substantially improves long-context generalization. Empirically, BAM enables accurate information retrieval at $500\times$ the training context length, outperforming previous state-of-the-art context length generalization in long context retrieval accuracy while maintaining comparable perplexity and introducing minimal additional parameters.
Authors: Vladimir Bataev, Andrei Andrusenko, Lilit Grigoryan, Aleksandr Laptev, Vitaly Lavrukhin, Boris Ginsburg
Abstract: Statistical n-gram language models are widely used for context-biasing tasks in Automatic Speech Recognition (ASR). However, existing implementations lack computational efficiency due to poor parallelization, making context-biasing less appealing for industrial use. This work rethinks data structures for statistical n-gram language models to enable fast and parallel operations for GPU-optimized inference. Our approach, named NGPU-LM, introduces customizable greedy decoding for all major ASR model types - including transducers, attention encoder-decoder models, and CTC - with less than 7% computational overhead. The proposed approach can eliminate more than 50% of the accuracy gap between greedy and beam search for out-of-domain scenarios while avoiding significant slowdown caused by beam search. The implementation of the proposed NGPU-LM is open-sourced.
Authors: Bargav Jayaraman, Virendra J. Marathe, Hamid Mozaffari, William F. Shen, Krishnaram Kenthapadi
Abstract: In enterprise settings, organizational data is segregated, siloed and carefully protected by elaborate access control frameworks. These access control structures can completely break down if an LLM fine-tuned on the siloed data serves requests, for downstream tasks, from individuals with disparate access privileges. We propose Permissioned LLMs (PermLLM), a new class of LLMs that superimpose the organizational data access control structures on query responses they generate. We formalize abstractions underpinning the means to determine whether access control enforcement happens correctly over LLM query responses. Our formalism introduces the notion of a relevant response that can be used to prove whether a PermLLM mechanism has been implemented correctly. We also introduce a novel metric, called access advantage, to empirically evaluate the efficacy of a PermLLM mechanism. We introduce three novel PermLLM mechanisms that build on Parameter Efficient Fine-Tuning to achieve the desired access control. We furthermore present two instantiations of access advantage--(i) Domain Distinguishability Index (DDI) based on Membership Inference Attacks, and (ii) Utility Gap Index (UGI) based on LLM utility evaluation. We demonstrate the efficacy of our PermLLM mechanisms through extensive experiments on four public datasets (GPQA, RCV1, SimpleQA, and WMDP), in addition to evaluating the validity of DDI and UGI metrics themselves for quantifying access control in LLMs.
Authors: Md Hasibul Amin, Mohammadreza Mohammadi, Jason D. Bakos, Ramtin Zand
Abstract: In this paper, we propose the CrossNAS framework, an automated approach for exploring a vast, multidimensional search space that spans various design abstraction layers-circuits, architecture, and systems-to optimize the deployment of machine learning workloads on analog processing-in-memory (PIM) systems. CrossNAS leverages the single-path one-shot weight-sharing strategy combined with the evolutionary search for the first time in the context of PIM system mapping and optimization. CrossNAS sets a new benchmark for PIM neural architecture search (NAS), outperforming previous methods in both accuracy and energy efficiency while maintaining comparable or shorter search times.
Authors: Junbo Yin, Chao Zha, Wenjia He, Chencheng Xu, Xin Gao
Abstract: Existing PLMs generate protein sequences based on a single-condition constraint from a specific modality, struggling to simultaneously satisfy multiple constraints across different modalities. In this work, we introduce CFP-Gen, a novel diffusion language model for Combinatorial Functional Protein GENeration. CFP-Gen facilitates the de novo protein design by integrating multimodal conditions with functional, sequence, and structural constraints. Specifically, an Annotation-Guided Feature Modulation (AGFM) module is introduced to dynamically adjust the protein feature distribution based on composable functional annotations, e.g., GO terms, IPR domains and EC numbers. Meanwhile, the Residue-Controlled Functional Encoding (RCFE) module captures residue-wise interaction to ensure more precise control. Additionally, off-the-shelf 3D structure encoders can be seamlessly integrated to impose geometric constraints. We demonstrate that CFP-Gen enables high-throughput generation of novel proteins with functionality comparable to natural proteins, while achieving a high success rate in designing multifunctional proteins. Code and data available at https://github.com/yinjunbo/cfpgen.
Authors: Maksim Kolodiazhnyi, Denis Tarasov, Dmitrii Zhemchuzhnikov, Alexander Nikulin, Ilya Zisman, Anna Vorontsova, Anton Konushin, Vladislav Kurenkov, Danila Rukhovich
Abstract: Computer-Aided Design (CAD) plays a central role in engineering and manufacturing, making it possible to create precise and editable 3D models. Using a variety of sensor or user-provided data as inputs for CAD reconstruction can democratize access to design applications. However, existing methods typically focus on a single input modality, such as point clouds, images, or text, which limits their generalizability and robustness. Leveraging recent advances in vision-language models (VLM), we propose a multi-modal CAD reconstruction model that simultaneously processes all three input modalities. Inspired by large language model (LLM) training paradigms, we adopt a two-stage pipeline: supervised fine-tuning (SFT) on large-scale procedurally generated data, followed by reinforcement learning (RL) fine-tuning using online feedback, obtained programatically. Furthermore, we are the first to explore RL fine-tuning of LLMs for CAD tasks demonstrating that online RL algorithms such as Group Relative Preference Optimization (GRPO) outperform offline alternatives. In the DeepCAD benchmark, our SFT model outperforms existing single-modal approaches in all three input modalities simultaneously. More importantly, after RL fine-tuning, cadrille sets new state-of-the-art on three challenging datasets, including a real-world one.
Authors: Haobo Zhang, Jiayu Zhou
Abstract: Fine-tuning large language models (LMs) for individual tasks yields strong performance but is expensive for deployment and storage. Recent works explore model merging to combine multiple task-specific models into a single multi-task model without additional training. However, existing merging methods often fail for models fine-tuned with low-rank adaptation (LoRA), due to significant performance degradation. In this paper, we show that this issue arises from a previously overlooked interplay between model parameters and data distributions. We propose Orthogonal Subspaces for Robust model Merging (OSRM) to constrain the LoRA subspace *prior* to fine-tuning, ensuring that updates relevant to one task do not adversely shift outputs for others. Our approach can seamlessly integrate with most existing merging algorithms, reducing the unintended interference among tasks. Extensive experiments on eight datasets, tested with three widely used LMs and two large LMs, demonstrate that our method not only boosts merging performance but also preserves single-task accuracy. Furthermore, our approach exhibits greater robustness to the hyperparameters of merging. These results highlight the importance of data-parameter interaction in model merging and offer a plug-and-play solution for merging LoRA models.
Authors: Niclas Boehmer, Sara Fish, Ariel D. Procaccia
Abstract: A key task in certain democratic processes is to produce a concise slate of statements that proportionally represents the full spectrum of user opinions. This task is similar to committee elections, but unlike traditional settings, the candidate set comprises all possible statements of varying lengths, and so it can only be accessed through specific queries. Combining social choice and large language models, prior work has approached this challenge through a framework of generative social choice. We extend the framework in two fundamental ways, providing theoretical guarantees even in the face of approximately optimal queries and a budget limit on the overall length of the slate. Using GPT-4o to implement queries, we showcase our approach on datasets related to city improvement measures and drug reviews, demonstrating its effectiveness in generating representative slates from unstructured user opinions.
Authors: Jaewoo Ahn, Heeseung Yun, Dayoon Ko, Gunhee Kim
Abstract: While pre-trained multimodal representations (e.g., CLIP) have shown impressive capabilities, they exhibit significant compositional vulnerabilities leading to counterintuitive judgments. We introduce Multimodal Adversarial Compositionality (MAC), a benchmark that leverages large language models (LLMs) to generate deceptive text samples to exploit these vulnerabilities across different modalities and evaluates them through both sample-wise attack success rate and group-wise entropy-based diversity. To improve zero-shot methods, we propose a self-training approach that leverages rejection-sampling fine-tuning with diversity-promoting filtering, which enhances both attack success rate and sample diversity. Using smaller language models like Llama-3.1-8B, our approach demonstrates superior performance in revealing compositional vulnerabilities across various multimodal representations, including images, videos, and audios.
Authors: Yuhui Zhang, Yuchang Su, Yiming Liu, Serena Yeung-Levy
Abstract: Negation is a fundamental linguistic phenomenon that can entirely reverse the meaning of a sentence. As vision language models (VLMs) continue to advance and are deployed in high-stakes applications, assessing their ability to comprehend negation becomes essential. To address this, we introduce NegVQA, a visual question answering (VQA) benchmark consisting of 7,379 two-choice questions covering diverse negation scenarios and image-question distributions. We construct NegVQA by leveraging large language models to generate negated versions of questions from existing VQA datasets. Evaluating 20 state-of-the-art VLMs across seven model families, we find that these models struggle significantly with negation, exhibiting a substantial performance drop compared to their responses to the original questions. Furthermore, we uncover a U-shaped scaling trend, where increasing model size initially degrades performance on NegVQA before leading to improvements. Our benchmark reveals critical gaps in VLMs' negation understanding and offers insights into future VLM development. Project page available at https://yuhui-zh15.github.io/NegVQA/.
Authors: Yongjin Yang, Euiin Yi, Jongwoo Ko, Kimin Lee, Zhijing Jin, Se-Young Yun
Abstract: The remarkable growth in large language model (LLM) capabilities has spurred exploration into multi-agent systems, with debate frameworks emerging as a promising avenue for enhanced problem-solving. These multi-agent debate (MAD) approaches, where agents collaboratively present, critique, and refine arguments, potentially offer improved reasoning, robustness, and diverse perspectives over monolithic models. Despite prior studies leveraging MAD, a systematic understanding of its effectiveness compared to self-agent methods, particularly under varying conditions, remains elusive. This paper seeks to fill this gap by conceptualizing MAD as a test-time computational scaling technique, distinguished by collaborative refinement and diverse exploration capabilities. We conduct a comprehensive empirical investigation comparing MAD with strong self-agent test-time scaling baselines on mathematical reasoning and safety-related tasks. Our study systematically examines the influence of task difficulty, model scale, and agent diversity on MAD's performance. Key findings reveal that, for mathematical reasoning, MAD offers limited advantages over self-agent scaling but becomes more effective with increased problem difficulty and decreased model capability, while agent diversity shows little benefit. Conversely, for safety tasks, MAD's collaborative refinement can increase vulnerability, but incorporating diverse agent configurations facilitates a gradual reduction in attack success through the collaborative refinement process. We believe our findings provide critical guidance for the future development of more effective and strategically deployed MAD systems.
Authors: Peixuan Han, Zijia Liu, Jiaxuan You
Abstract: Large language models (LLMs) have shown promising potential in persuasion, but existing works on training LLM persuaders are still preliminary. Notably, while humans are skilled in modeling their opponent's thoughts and opinions proactively and dynamically, current LLMs struggle with such Theory of Mind (ToM) reasoning, resulting in limited diversity and opponent awareness. To address this limitation, we introduce Theory of Mind Augmented Persuader (ToMAP), a novel approach for building more flexible persuader agents by incorporating two theory of mind modules that enhance the persuader's awareness and analysis of the opponent's mental state. Specifically, we begin by prompting the persuader to consider possible objections to the target central claim, and then use a text encoder paired with a trained MLP classifier to predict the opponent's current stance on these counterclaims. Our carefully designed reinforcement learning schema enables the persuader learns how to analyze opponent-related information and utilize it to generate more effective arguments. Experiments show that the ToMAP persuader, while containing only 3B parameters, outperforms much larger baselines, like GPT-4o, with a relative gain of 39.4% across multiple persuadee models and diverse corpora. Notably, ToMAP exhibits complex reasoning chains and reduced repetition during training, which leads to more diverse and effective arguments. The opponent-aware feature of ToMAP also makes it suitable for long conversations and enables it to employ more logical and opponent-aware strategies. These results underscore our method's effectiveness and highlight its potential for developing more persuasive language agents. Code is available at: https://github.com/ulab-uiuc/ToMAP.
Authors: Sheng Zhang, Qin Liu, Naoto Usuyama, Cliff Wong, Tristan Naumann, Hoifung Poon
Abstract: The emergence of scaling laws has profoundly shaped the development of large language models (LLMs), enabling predictable performance gains through systematic increases in model size, dataset volume, and compute. Yet, these principles remain largely unexplored in the context of electronic health records (EHRs) -- a rich, sequential, and globally abundant data source that differs structurally from natural language. In this work, we present the first empirical investigation of scaling laws for EHR foundation models. By training transformer architectures on patient timeline data from the MIMIC-IV database across varying model sizes and compute budgets, we identify consistent scaling patterns, including parabolic IsoFLOPs curves and power-law relationships between compute, model parameters, data size, and clinical utility. These findings demonstrate that EHR models exhibit scaling behavior analogous to LLMs, offering predictive insights into resource-efficient training strategies. Our results lay the groundwork for developing powerful EHR foundation models capable of transforming clinical prediction tasks and advancing personalized healthcare.
Authors: Yuntao Ma, Andrei Cramariuc, Farbod Farshidian, Marco Hutter
Abstract: Coordinating the motion between lower and upper limbs and aligning limb control with perception are substantial challenges in robotics, particularly in dynamic environments. To this end, we introduce an approach for enabling legged mobile manipulators to play badminton, a task that requires precise coordination of perception, locomotion, and arm swinging. We propose a unified reinforcement learning-based control policy for whole-body visuomotor skills involving all degrees of freedom to achieve effective shuttlecock tracking and striking. This policy is informed by a perception noise model that utilizes real-world camera data, allowing for consistent perception error levels between simulation and deployment and encouraging learned active perception behaviors. Our method includes a shuttlecock prediction model, constrained reinforcement learning for robust motion control, and integrated system identification techniques to enhance deployment readiness. Extensive experimental results in a variety of environments validate the robot's capability to predict shuttlecock trajectories, navigate the service area effectively, and execute precise strikes against human players, demonstrating the feasibility of using legged mobile manipulators in complex and dynamic sports scenarios.
Authors: Pin-Han Chen, Yu-Sheng Lin, Wei-Cheng Lee, Tin-Yu Leu, Po-Hsiang Hsu, Anjana Dissanayake, Sungjin Oh, Chinq-Shiun Chiu
Abstract: RF/Analog design is essential for bridging digital technologies with real-world signals, ensuring the functionality and reliability of a wide range of electronic systems. However, analog design procedures are often intricate, time-consuming and reliant on expert intuition, and hinder the time and cost efficiency of circuit development. To overcome the limitations of the manual circuit design, we introduce MenTeR - a multiagent workflow integrated into an end-to-end analog design framework. By employing multiple specialized AI agents that collaboratively address different aspects of the design process, such as specification understanding, circuit optimization, and test bench validation, MenTeR reduces the dependency on frequent trial-and-error-style intervention. MenTeR not only accelerates the design cycle time but also facilitates a broader exploration of the design space, demonstrating robust capabilities in handling real-world analog systems. We believe that MenTeR lays the groundwork for future "RF/Analog Copilots" that can collaborate seamlessly with human designers.
Authors: Agnideep Aich, Ashit Baran Aich, Bruce Wade
Abstract: Traditional classifiers often assume feature independence or rely on overly simplistic relationships, leading to poor performance in settings where real-world dependencies matter. We introduce the Deep Copula Classifier (DCC), a generative model that separates the learning of each feature's marginal distribution from the modeling of their joint dependence structure via neural network-parameterized copulas. For each class, lightweight neural networks are used to flexibly and adaptively capture feature interactions, making DCC particularly effective when classification is driven by complex dependencies. We establish that DCC converges to the Bayes-optimal classifier under standard conditions and provide explicit convergence rates of O(n^{-r/(2r + d)}) for r-smooth copula densities. Beyond theoretical guarantees, we outline several practical extensions, including high-dimensional scalability through vine and factor copula architectures, semi-supervised learning via entropy regularization, and online adaptation using streaming gradient methods. By unifying statistical rigor with the representational power of neural networks, DCC offers a mathematically grounded and interpretable framework for dependency-aware classification.
Authors: Guangyi Liu, Yongqi Zhang, Xunyuan Liu, Quanming Yao
Abstract: Drug-drug interaction (DDI) prediction is critical for treatment safety. While large language models (LLMs) show promise in pharmaceutical tasks, their effectiveness in DDI prediction remains challenging. Inspired by the well-established clinical practice where physicians routinely reference similar historical cases to guide their decisions through case-based reasoning (CBR), we propose CBR-DDI, a novel framework that distills pharmacological principles from historical cases to improve LLM reasoning for DDI tasks. CBR-DDI constructs a knowledge repository by leveraging LLMs to extract pharmacological insights and graph neural networks (GNNs) to model drug associations. A hybrid retrieval mechanism and dual-layer knowledge-enhanced prompting allow LLMs to effectively retrieve and reuse relevant cases. We further introduce a representative sampling strategy for dynamic case refinement. Extensive experiments demonstrate that CBR-DDI achieves state-of-the-art performance, with a significant 28.7% accuracy improvement over both popular LLMs and CBR baseline, while maintaining high interpretability and flexibility.
Authors: Zijian Liu, Zhengyuan Zhou
Abstract: We study the convergence of the shuffling gradient method, a popular algorithm employed to minimize the finite-sum function with regularization, in which functions are passed to apply (Proximal) Gradient Descent (GD) one by one whose order is determined by a permutation on the indices of functions. In contrast to its easy implementation and effective performance in practice, the theoretical understanding remains limited. A recent advance by (Liu & Zhou, 2024b) establishes the first last-iterate convergence results under various settings, especially proving the optimal rates for smooth (strongly) convex optimization. However, their bounds for nonsmooth (strongly) convex functions are only as fast as Proximal GD. In this work, we provide the first improved last-iterate analysis for the nonsmooth case demonstrating that the widely used Random Reshuffle ($\textsf{RR}$) and Single Shuffle ($\textsf{SS}$) strategies are both provably faster than Proximal GD, reflecting the benefit of randomness. As an important implication, we give the first (nearly) optimal convergence result for the suffix average under the $\textsf{RR}$ sampling scheme in the general convex case, matching the lower bound shown by (Koren et al., 2022).
Authors: Shuyin Xia, Xiaojiang Tian, Suzhen Yuan, Jeremiah D. Deng
Abstract: High time complexity is one of the biggest challenges faced by $k$-Nearest Neighbors ($k$NN). Although current classical and quantum $k$NN algorithms have made some improvements, they still have a speed bottleneck when facing large amounts of data. To address this issue, we propose an innovative algorithm called Granular-Ball based Quantum $k$NN(GB-Q$k$NN). This approach achieves higher efficiency by first employing granular-balls, which reduces the data size needed to processed. The search process is then accelerated by adopting a Hierarchical Navigable Small World (HNSW) method. Moreover, we optimize the time-consuming steps, such as distance calculation, of the HNSW via quantization, further reducing the time complexity of the construct and search process. By combining the use of granular-balls and quantization of the HNSW method, our approach manages to take advantage of these treatments and significantly reduces the time complexity of the $k$NN-like algorithms, as revealed by a comprehensive complexity analysis.
Authors: Amit Kumthekar, Zion Tilley, Henry Duong, Bhargav Patel, Michael Magnoli, Ahmed Omar, Ahmed Nasser, Chaitanya Gharpure, Yevgen Reztzov
Abstract: Despite the growing clinical adoption of large language models (LLMs), current approaches heavily rely on single model architectures. To overcome risks of obsolescence and rigid dependence on single model systems, we present a novel framework, termed the Consensus Mechanism. Mimicking clinical triage and multidisciplinary clinical decision-making, the Consensus Mechanism implements an ensemble of specialized medical expert agents enabling improved clinical decision making while maintaining robust adaptability. This architecture enables the Consensus Mechanism to be optimized for cost, latency, or performance, purely based on its interior model configuration. To rigorously evaluate the Consensus Mechanism, we employed three medical evaluation benchmarks: MedMCQA, MedQA, and MedXpertQA Text, and the differential diagnosis dataset, DDX+. On MedXpertQA, the Consensus Mechanism achieved an accuracy of 61.0% compared to 53.5% and 45.9% for OpenAI's O3 and Google's Gemini 2.5 Pro. Improvement was consistent across benchmarks with an increase in accuracy on MedQA ($\Delta\mathrm{Accuracy}_{\mathrm{consensus\text{-}O3}} = 3.4\%$) and MedMCQA ($\Delta\mathrm{Accuracy}_{\mathrm{consensus\text{-}O3}} = 9.1\%$). These accuracy gains extended to differential diagnosis generation, where our system demonstrated improved recall and precision (F1$_\mathrm{consensus}$ = 0.326 vs. F1$_{\mathrm{O3\text{-}high}}$ = 0.2886) and a higher top-1 accuracy for DDX (Top1$_\mathrm{consensus}$ = 52.0% vs. Top1$_{\mathrm{O3\text{-}high}}$ = 45.2%).
Authors: Wenzhi Gao, Ya-Chi Chu, Yinyu Ye, Madeleine Udell
Abstract: This paper establishes the theoretical foundations of the online scaled gradient methods (OSGM), a framework that utilizes online learning to adapt stepsizes and provably accelerate first-order methods. OSGM quantifies the effectiveness of a stepsize by a feedback function motivated from a convergence measure and uses the feedback to adjust the stepsize through an online learning algorithm. Consequently, instantiations of OSGM achieve convergence rates that are asymptotically no worse than the optimal stepsize. OSGM yields desirable convergence guarantees on smooth convex problems, including 1) trajectory-dependent global convergence on smooth convex objectives; 2) an improved complexity result on smooth strongly convex problems, and 3) local superlinear convergence. Notably, OSGM constitutes a new family of first-order methods with non-asymptotic superlinear convergence, joining the celebrated quasi-Newton methods. Finally, OSGM explains the empirical success of the popular hypergradient-descent heuristic in optimization for machine learning.
Authors: Junyan Liu, Arnab Maiti, Artin Tajdini, Kevin Jamieson, Lillian J. Ratliff
Abstract: We initiate the study of a repeated principal-agent problem over a finite horizon $T$, where a principal sequentially interacts with $K\geq 2$ types of agents arriving in an adversarial order. At each round, the principal strategically chooses one of the $N$ arms to incentivize for an arriving agent of unknown type. The agent then chooses an arm based on its own utility and the provided incentive, and the principal receives a corresponding reward. The objective is to minimize regret against the best incentive in hindsight. Without prior knowledge of agent behavior, we show that the problem becomes intractable, leading to linear regret. We analyze two key settings where sublinear regret is achievable. In the first setting, the principal knows the arm each agent type would select greedily for any given incentive. Under this setting, we propose an algorithm that achieves a regret bound of $O(\min\{\sqrt{KT\log N},K\sqrt{T}\})$ and provide a matching lower bound up to a $\log K$ factor. In the second setting, an agent's response varies smoothly with the incentive and is governed by a Lipschitz constant $L\geq 1$. Under this setting, we show that there is an algorithm with a regret bound of $\tilde{O}((LN)^{1/3}T^{2/3})$ and establish a matching lower bound up to logarithmic factors. Finally, we extend our algorithmic results for both settings by allowing the principal to incentivize multiple arms simultaneously in each round.
Authors: Jeongsol Kim, Yeobin Hong, Jong Chul Ye
Abstract: Recent inversion-free, flow-based image editing methods such as FlowEdit leverages a pre-trained noise-to-image flow model such as Stable Diffusion 3, enabling text-driven manipulation by solving an ordinary differential equation (ODE). While the lack of exact latent inversion is a core advantage of these methods, it often results in unstable editing trajectories and poor source consistency. To address this limitation, we propose FlowAlign, a novel inversion-free flow-based framework for consistent image editing with principled trajectory control. FlowAlign introduces a flow-matching loss as a regularization mechanism to promote smoother and more stable trajectories during the editing process. Notably, the flow-matching loss is shown to explicitly balance semantic alignment with the edit prompt and structural consistency with the source image along the trajectory. Furthermore, FlowAlign naturally supports reverse editing by simply reversing the ODE trajectory, highlighting the reversible and consistent nature of the transformation. Extensive experiments demonstrate that FlowAlign outperforms existing methods in both source preservation and editing controllability.
Authors: Lorenzo Marinucci, Claudio Battiloro, Paolo Di Lorenzo
Abstract: This paper introduces a novel adaptive framework for processing dynamic flow signals over simplicial complexes, extending classical least-mean-squares (LMS) methods to high-order topological domains. Building on discrete Hodge theory, we present a topological LMS algorithm that efficiently processes streaming signals observed over time-varying edge subsets. We provide a detailed stochastic analysis of the algorithm, deriving its stability conditions, steady-state mean-square-error, and convergence speed, while exploring the impact of edge sampling on performance. We also propose strategies to design optimal edge sampling probabilities, minimizing rate while ensuring desired estimation accuracy. Assuming partial knowledge of the complex structure (e.g., the underlying graph), we introduce an adaptive topology inference method that integrates with the proposed LMS framework. Additionally, we propose a distributed version of the algorithm and analyze its stability and mean-square-error properties. Empirical results on synthetic and real-world traffic data demonstrate that our approach, in both centralized and distributed settings, outperforms graph-based LMS methods by leveraging higher-order topological features.
Authors: Antonio D'Orazio, Maria Rosaria Briglia, Donato Crisostomi, Dario Loi, Emanuele Rodol\`a, Iacopo Masi
Abstract: CLIP is a discriminative model trained to align images and text in a shared embedding space. Due to its multimodal structure, it serves as the backbone of many generative pipelines, where a decoder is trained to map from the shared space back to images. In this work, we show that image synthesis is nevertheless possible using CLIP alone -- without any decoder, training, or fine-tuning. Our approach optimizes a frequency-aware implicit neural representation that encourages coarse-to-fine generation by stratifying frequencies across network layers. To stabilize this inverse mapping, we introduce adversarially robust initialization, a lightweight Orthogonal Procrustes projection to align local text and image embeddings, and a blending loss that anchors outputs to natural image statistics. Without altering CLIP's weights, this framework unlocks capabilities such as text-to-image generation, style transfer, and image reconstruction. These findings suggest that discriminative models may hold untapped generative potential, hidden in plain sight.
Authors: Eshant English, Christoph Lippert
Abstract: Conformal prediction provides a model-agnostic framework for uncertainty quantification with finite-sample validity guarantees, making it an attractive tool for constructing reliable prediction sets. However, existing approaches commonly rely on residual-based conformity scores, which impose geometric constraints and struggle when the underlying distribution is multimodal. In particular, they tend to produce overly conservative prediction areas centred around the mean, often failing to capture the true shape of complex predictive distributions. In this work, we introduce JAPAN (Joint Adaptive Prediction Areas with Normalising-Flows), a conformal prediction framework that uses density-based conformity scores. By leveraging flow-based models, JAPAN estimates the (predictive) density and constructs prediction areas by thresholding on the estimated density scores, enabling compact, potentially disjoint, and context-adaptive regions that retain finite-sample coverage guarantees. We theoretically motivate the efficiency of JAPAN and empirically validate it across multivariate regression and forecasting tasks, demonstrating good calibration and tighter prediction areas compared to existing baselines. We also provide several \emph{extensions} adding flexibility to our proposed framework.
Authors: Zhaokai Sun, Li Zhang, Qing Wang, Pan Zhou, Lei Xie
Abstract: Overlapping Speech Detection (OSD) aims to identify regions where multiple speakers overlap in a conversation, a critical challenge in multi-party speech processing. This work proposes a speaker-aware progressive OSD model that leverages a progressive training strategy to enhance the correlation between subtasks such as voice activity detection (VAD) and overlap detection. To improve acoustic representation, we explore the effectiveness of state-of-the-art self-supervised learning (SSL) models, including WavLM and wav2vec 2.0, while incorporating a speaker attention module to enrich features with frame-level speaker information. Experimental results show that the proposed method achieves state-of-the-art performance, with an F1 score of 82.76\% on the AMI test set, demonstrating its robustness and effectiveness in OSD.
Authors: T. Jahn, J. Chemseddine, P. Hagemann, C. Wald, G. Steidl
Abstract: Accurately modeling time-continuous stochastic processes from irregular observations remains a significant challenge. In this paper, we leverage ideas from generative modeling of image data to push the boundary of time series generation. For this, we find new generators of SDEs and jump processes, inspired by trajectory flow matching, that have the marginal distributions of the time series of interest. Specifically, we can handle discontinuities of the underlying processes by parameterizing the jump kernel densities by scaled Gaussians that allow for closed form formulas of the corresponding Kullback-Leibler divergence in the loss. Unlike most other approaches, we are able to handle irregularly sampled time series.
Authors: Hemant Tyagi
Abstract: Given an undirected and connected graph $G$ on $T$ vertices, suppose each vertex $t$ has a latent signal $x_t \in \mathbb{R}^n$ associated to it. Given partial linear measurements of the signals, for a potentially small subset of the vertices, our goal is to estimate $x_t$'s. Assuming that the signals are smooth w.r.t $G$, in the sense that the quadratic variation of the signals over the graph is small, we obtain non-asymptotic bounds on the mean squared error for jointly recovering $x_t$'s, for the smoothness penalized least squares estimator. In particular, this implies for certain choices of $G$ that this estimator is weakly consistent (as $T \rightarrow \infty$) under potentially very stringent sampling, where only one coordinate is measured per vertex for a vanishingly small fraction of the vertices. The results are extended to a ``multi-layer'' ranking problem where $x_t$ corresponds to the latent strengths of a collection of $n$ items, and noisy pairwise difference measurements are obtained at each ``layer'' $t$ via a measurement graph $G_t$. Weak consistency is established for certain choices of $G$ even when the individual $G_t$'s are very sparse and disconnected.
Authors: Budhaditya Halder, Shubhayan Pan, Koulik Khamaru
Abstract: We consider the problem of statistical inference when the data is collected via a Thompson Sampling-type algorithm. While Thompson Sampling (TS) is known to be both asymptotically optimal and empirically effective, its adaptive sampling scheme poses challenges for constructing confidence intervals for model parameters. We propose and analyze a variant of TS, called Stable Thompson Sampling, in which the posterior variance is inflated by a logarithmic factor. We show that this modification leads to asymptotically normal estimates of the arm means, despite the non-i.i.d. nature of the data. Importantly, this statistical benefit comes at a modest cost: the variance inflation increases regret by only a logarithmic factor compared to standard TS. Our results reveal a principled trade-off: by paying a small price in regret, one can enable valid statistical inference for adaptive decision-making algorithms.
Authors: Mao-Lin Luo, Zi-Hao Zhou, Tong Wei, Min-Ling Zhang
Abstract: Continual learning with vision-language models like CLIP offers a pathway toward scalable machine learning systems by leveraging its transferable representations. Existing CLIP-based methods adapt the pre-trained image encoder by adding multiple sets of learnable parameters, with each task using a partial set of parameters. This requires selecting the expected parameters for input images during inference, which is prone to error that degrades performance. To address this problem, we introduce LADA (Label-specific ADApter). Instead of partitioning parameters across tasks, LADA appends lightweight, label-specific memory units to the frozen CLIP image encoder, enabling discriminative feature generation by aggregating task-agnostic knowledge. To prevent catastrophic forgetting, LADA employs feature distillation for seen classes, preventing their features from being interfered with by new classes. Positioned after the image encoder, LADA prevents gradient flow to the frozen CLIP parameters, ensuring efficient training. Extensive results show that LADA achieves state-of-the-art performance in continual learning settings. The implementation code is available at https://github.com/MaolinLuo/LADA.
Authors: James Xu Zhao, Jimmy Z. J. Liu, Bryan Hooi, See-Kiong Ng
Abstract: Large language models (LLMs) are widely used for long-form text generation. However, factual errors in the responses would undermine their reliability. Despite growing attention to LLM factuality, the effect of response length on factuality remains underexplored. In this work, we systematically investigate this relationship by first introducing an automatic and bi-level long-form factuality evaluation framework, which achieves high agreement with human annotations while being cost-effective. Using this framework, we conduct controlled experiments and find that longer responses exhibit lower factual precision, confirming the presence of length bias. To explain this phenomenon, we empirically examine three hypotheses: error propagation, long context, and facts exhaustion. Our results reveal that facts exhaustion, where the model gradually exhausts more reliable knowledge, is the primary cause of factual degradation, rather than the other two hypotheses.
Authors: Yunkee Chae, Kyogu Lee
Abstract: We present MGE-LDM, a unified latent diffusion framework for simultaneous music generation, source imputation, and query-driven source separation. Unlike prior approaches constrained to fixed instrument classes, MGE-LDM learns a joint distribution over full mixtures, submixtures, and individual stems within a single compact latent diffusion model. At inference, MGE-LDM enables (1) complete mixture generation, (2) partial generation (i.e., source imputation), and (3) text-conditioned extraction of arbitrary sources. By formulating both separation and imputation as conditional inpainting tasks in the latent space, our approach supports flexible, class-agnostic manipulation of arbitrary instrument sources. Notably, MGE-LDM can be trained jointly across heterogeneous multi-track datasets (e.g., Slakh2100, MUSDB18, MoisesDB) without relying on predefined instrument categories. Audio samples are available at our project page: https://yoongi43.github.io/MGELDM_Samples/.
Authors: Weizhe Kong, Xiao Wang, Ruichong Gao, Chenglong Li, Yu Zhang, Xing Yang, Yaowei Wang, Jin Tang
Abstract: Pedestrian Attribute Recognition (PAR) is an indispensable task in human-centered research and has made great progress in recent years with the development of deep neural networks. However, the potential vulnerability and anti-interference ability have still not been fully explored. To bridge this gap, this paper proposes the first adversarial attack and defense framework for pedestrian attribute recognition. Specifically, we exploit both global- and patch-level attacks on the pedestrian images, based on the pre-trained CLIP-based PAR framework. It first divides the input pedestrian image into non-overlapping patches and embeds them into feature embeddings using a projection layer. Meanwhile, the attribute set is expanded into sentences using prompts and embedded into attribute features using a pre-trained CLIP text encoder. A multi-modal Transformer is adopted to fuse the obtained vision and text tokens, and a feed-forward network is utilized for attribute recognition. Based on the aforementioned PAR framework, we adopt the adversarial semantic and label-perturbation to generate the adversarial noise, termed ASL-PAR. We also design a semantic offset defense strategy to suppress the influence of adversarial attacks. Extensive experiments conducted on both digital domains (i.e., PETA, PA100K, MSP60K, RAPv2) and physical domains fully validated the effectiveness of our proposed adversarial attack and defense strategies for the pedestrian attribute recognition. The source code of this paper will be released on https://github.com/Event-AHU/OpenPAR.
Authors: Abhirup Chakravarty, Mark Brenchley, Trevor Breakspear, Ian Lewin, Yan Huang
Abstract: A key ethical challenge in Automated Essay Scoring (AES) is ensuring that scores are only released when they meet high reliability standards. Confidence modelling addresses this by assigning a reliability estimate measure, in the form of a confidence score, to each automated score. In this study, we frame confidence estimation as a classification task: predicting whether an AES-generated score correctly places a candidate in the appropriate CEFR level. While this is a binary decision, we leverage the inherent granularity of the scoring domain in two ways. First, we reformulate the task as an n-ary classification problem using score binning. Second, we introduce a set of novel Kernel Weighted Ordinal Categorical Cross Entropy (KWOCCE) loss functions that incorporate the ordinal structure of CEFR labels. Our best-performing model achieves an F1 score of 0.97, and enables the system to release 47% of scores with 100% CEFR agreement and 99% with at least 95% CEFR agreement -compared to approximately 92% (approx.) CEFR agreement from the standalone AES model where we release all AM predicted scores.
Authors: Matteo Gallici, Haitz S\'aez de Oc\'ariz Borde
Abstract: Fine-tuning pre-trained generative models with Reinforcement Learning (RL) has emerged as an effective approach for aligning outputs more closely with nuanced human preferences. In this paper, we investigate the application of Group Relative Policy Optimization (GRPO) to fine-tune next-scale visual autoregressive (VAR) models. Our empirical results demonstrate that this approach enables alignment to intricate reward signals derived from aesthetic predictors and CLIP embeddings, significantly enhancing image quality and enabling precise control over the generation style. Interestingly, by leveraging CLIP, our method can help VAR models generalize beyond their initial ImageNet distribution: through RL-driven exploration, these models can generate images aligned with prompts referencing image styles that were absent during pre-training. In summary, we show that RL-based fine-tuning is both efficient and effective for VAR models, benefiting particularly from their fast inference speeds, which are advantageous for online sampling, an aspect that poses significant challenges for diffusion-based alternatives.
Authors: Jakub Martinka, Lina Zhang, Yi-Fan Hou, Miko{\l}aj Martyka, Ji\v{r}\'i Pittner, Mario Barbatti, Pavlo O. Dral
Abstract: Nonadiabatic couplings (NACs) play a crucial role in modeling photochemical and photophysical processes with methods such as the widely used fewest-switches surface hopping (FSSH). There is therefore a strong incentive to machine learn NACs for accelerating simulations. However, this is challenging due to NACs' vectorial, double-valued character and the singularity near a conical intersection seam. For the first time, we design NAC-specific descriptors based on our domain expertise and show that they allow learning NACs with never-before-reported accuracy of $R^2$ exceeding 0.99. The key to success is also our new ML phase-correction procedure. We demonstrate the efficiency and robustness of our approach on a prototypical example of fully ML-driven FSSH simulations of fulvene targeting the SA-2-CASSCF(6,6) electronic structure level. This ML-FSSH dynamics leads to an accurate description of $S_1$ decay while reducing error bars by allowing the execution of a large ensemble of trajectories. Our implementations are available in open-source MLatom.
Authors: Ivana Nikoloska, Hamdi Joudeh, Ruud van Sloun, Osvaldo Simeone
Abstract: Quantum sensing exploits non-classical effects to overcome limitations of classical sensors, with applications ranging from gravitational-wave detection to nanoscale imaging. However, practical quantum sensors built on noisy intermediate-scale quantum (NISQ) devices face significant noise and sampling constraints, and current variational quantum sensing (VQS) methods lack rigorous performance guarantees. This paper proposes an online control framework for VQS that dynamically updates the variational parameters while providing deterministic error bars on the estimates. By leveraging online conformal inference techniques, the approach produces sequential estimation sets with a guaranteed long-term risk level. Experiments on a quantum magnetometry task confirm that the proposed dynamic VQS approach maintains the required reliability over time, while still yielding precise estimates. The results demonstrate the practical benefits of combining variational quantum algorithms with online conformal inference to achieve reliable quantum sensing on NISQ devices.
Authors: Binyamin Manela, Sharon Gannot, Ethan Fetyaya
Abstract: Visual dubbing, the synchronization of facial movements with new speech, is crucial for making content accessible across different languages, enabling broader global reach. However, current methods face significant limitations. Existing approaches often generate talking faces, hindering seamless integration into original scenes, or employ inpainting techniques that discard vital visual information like partial occlusions and lighting variations. This work introduces EdiDub, a novel framework that reformulates visual dubbing as a content-aware editing task. EdiDub preserves the original video context by utilizing a specialized conditioning scheme to ensure faithful and accurate modifications rather than mere copying. On multiple benchmarks, including a challenging occluded-lip dataset, EdiDub significantly improves identity preservation and synchronization. Human evaluations further confirm its superiority, achieving higher synchronization and visual naturalness scores compared to the leading methods. These results demonstrate that our content-aware editing approach outperforms traditional generation or inpainting, particularly in maintaining complex visual elements while ensuring accurate lip synchronization.
Authors: Jang-Hyun Kim, Jinuk Kim, Sangwoo Kwon, Jae W. Lee, Sangdoo Yun, Hyun Oh Song
Abstract: Transformer-based large language models (LLMs) cache context as key-value (KV) pairs during inference. As context length grows, KV cache sizes expand, leading to substantial memory overhead and increased attention latency. This paper introduces KVzip, a query-agnostic KV cache eviction method enabling effective reuse of compressed KV caches across diverse queries. KVzip quantifies the importance of a KV pair using the underlying LLM to reconstruct original contexts from cached KV pairs, subsequently evicting pairs with lower importance. Extensive empirical evaluations demonstrate that KVzip reduces KV cache size by 3-4$\times$ and FlashAttention decoding latency by approximately 2$\times$, with negligible performance loss in question-answering, retrieval, reasoning, and code comprehension tasks. Evaluations include various models such as LLaMA3.1-8B, Qwen2.5-14B, and Gemma3-12B, with context lengths reaching up to 170K tokens. KVzip significantly outperforms existing query-aware KV eviction methods, which suffer from performance degradation even at a 90% cache budget ratio under multi-query scenarios.
Authors: Amer Krivo\v{s}ija, Alexander Munteanu, Andr\'e Nusser, Chris Schwiegelshohn
Abstract: This paper introduces $k$-Dynamic Time Warping ($k$-DTW), a novel dissimilarity measure for polygonal curves. $k$-DTW has stronger metric properties than Dynamic Time Warping (DTW) and is more robust to outliers than the Fr\'{e}chet distance, which are the two gold standards of dissimilarity measures for polygonal curves. We show interesting properties of $k$-DTW and give an exact algorithm as well as a $(1+\varepsilon)$-approximation algorithm for $k$-DTW by a parametric search for the $k$-th largest matched distance. We prove the first dimension-free learning bounds for curves and further learning theoretic results. $k$-DTW not only admits smaller sample size than DTW for the problem of learning the median of curves, where some factors depending on the curves' complexity $m$ are replaced by $k$, but we also show a surprising separation on the associated Rademacher and Gaussian complexities: $k$-DTW admits strictly smaller bounds than DTW, by a factor $\tilde\Omega(\sqrt{m})$ when $k\ll m$. We complement our theoretical findings with an experimental illustration of the benefits of using $k$-DTW for clustering and nearest neighbor classification.
Authors: Daniel Jarne Ornia, Nicholas Bishop, Joel Dyer, Wei-Chen Lee, Ani Calinescu, Doyne Farme, Michael Wooldridge
Abstract: Advanced reasoning models with agentic capabilities (AI agents) are deployed to interact with humans and to solve sequential decision-making problems under (approximate) utility functions and internal models. When such problems have resource or failure constraints where action sequences may be forcibly terminated once resources are exhausted, agents face implicit trade-offs that reshape their utility-driven (rational) behaviour. Additionally, since these agents are typically commissioned by a human principal to act on their behalf, asymmetries in constraint exposure can give rise to previously unanticipated misalignment between human objectives and agent incentives. We formalise this setting through a survival bandit framework, provide theoretical and empirical results that quantify the impact of survival-driven preference shifts, identify conditions under which misalignment emerges and propose mechanisms to mitigate the emergence of risk-seeking or risk-averse behaviours. As a result, this work aims to increase understanding and interpretability of emergent behaviours of AI agents operating under such survival pressure, and offer guidelines for safely deploying such AI systems in critical resource-limited environments.
Authors: Adrien Majka, El-Mahdi El-Mhamdi
Abstract: Goodhart's law is a famous adage in policy-making that states that ``When a measure becomes a target, it ceases to be a good measure''. As machine learning models and the optimisation capacity to train them grow, growing empirical evidence reinforced the belief in the validity of this law without however being formalised. Recently, a few attempts were made to formalise Goodhart's law, either by categorising variants of it, or by looking at how optimising a proxy metric affects the optimisation of an intended goal. In this work, we alleviate the simplifying independence assumption, made in previous works, and the assumption on the learning paradigm made in most of them, to study the effect of the coupling between the proxy metric and the intended goal on Goodhart's law. Our results show that in the case of light tailed goal and light tailed discrepancy, dependence does not change the nature of Goodhart's effect. However, in the light tailed goal and heavy tailed discrepancy case, we exhibit an example where over-optimisation occurs at a rate inversely proportional to the heavy tailedness of the discrepancy between the goal and the metric. %
Authors: Ron Shapira Weber, Shahar Ben Ishay, Andrey Lavrinenko, Shahaf E. Finder, Oren Freifeld
Abstract: Fast and scalable alignment of time series is a fundamental challenge in many domains. The standard solution, Dynamic Time Warping (DTW), struggles with poor scalability and sensitivity to noise. We introduce TimePoint, a self-supervised method that dramatically accelerates DTW-based alignment while typically improving alignment accuracy by learning keypoints and descriptors from synthetic data. Inspired by 2D keypoint detection but carefully adapted to the unique challenges of 1D signals, TimePoint leverages efficient 1D diffeomorphisms, which effectively model nonlinear time warping, to generate realistic training data. This approach, along with fully convolutional and wavelet convolutional architectures, enables the extraction of informative keypoints and descriptors. Applying DTW to these sparse representations yield major speedups and typically higher alignment accuracy than standard DTW applied to the full signals. TimePoint demonstrates strong generalization to real-world time series when trained solely on synthetic data, and further improves with fine-tuning on real data. Extensive experiments demonstrate that TimePoint consistently achieves faster and more accurate alignments than standard DTW, making it a scalable solution for time-series analysis. Our code is available at https://github.com/BGU-CS-VIL/TimePoint
Authors: Liangliang Zhang, Zhuorui Jiang, Hongliang Chi, Haoyang Chen, Mohammed Elkoumy, Fali Wang, Qiong Wu, Zhengyi Zhou, Shirui Pan, Suhang Wang, Yao Ma
Abstract: Knowledge Graph Question Answering (KGQA) systems rely on high-quality benchmarks to evaluate complex multi-hop reasoning. However, despite their widespread use, popular datasets such as WebQSP and CWQ suffer from critical quality issues, including inaccurate or incomplete ground-truth annotations, poorly constructed questions that are ambiguous, trivial, or unanswerable, and outdated or inconsistent knowledge. Through a manual audit of 16 popular KGQA datasets, including WebQSP and CWQ, we find that the average factual correctness rate is only 57 %. To address these issues, we introduce KGQAGen, an LLM-in-the-loop framework that systematically resolves these pitfalls. KGQAGen combines structured knowledge grounding, LLM-guided generation, and symbolic verification to produce challenging and verifiable QA instances. Using KGQAGen, we construct KGQAGen-10k, a ten-thousand scale benchmark grounded in Wikidata, and evaluate a diverse set of KG-RAG models. Experimental results demonstrate that even state-of-the-art systems struggle on this benchmark, highlighting its ability to expose limitations of existing models. Our findings advocate for more rigorous benchmark construction and position KGQAGen as a scalable framework for advancing KGQA evaluation.
Authors: Andrew Chang, Yike Li, Iran R. Roman, David Poeppel
Abstract: Audio DNNs have demonstrated impressive performance on various machine listening tasks; however, most of their representations are computationally costly and uninterpretable, leaving room for optimization. Here, we propose a novel approach centered on spectrotemporal modulation (STM) features, a signal processing method that mimics the neurophysiological representation in the human auditory cortex. The classification performance of our STM-based model, without any pretraining, is comparable to that of pretrained audio DNNs across diverse naturalistic speech, music, and environmental sounds, which are essential categories for both human cognition and machine perception. These results show that STM is an efficient and interpretable feature representation for audio classification, advancing the development of machine listening and unlocking exciting new possibilities for basic understanding of speech and auditory sciences, as well as developing audio BCI and cognitive computing.
Authors: Sanberk Serbest, Tijana Stojkovic, Milos Cernak, Andrew Harper
Abstract: In this work, we propose a full-band real-time speech enhancement system with GAN-based stochastic regeneration. Predictive models focus on estimating the mean of the target distribution, whereas generative models aim to learn the full distribution. This behavior of predictive models may lead to over-suppression, i.e. the removal of speech content. In the literature, it was shown that combining a predictive model with a generative one within the stochastic regeneration framework can reduce the distortion in the output. We use this framework to obtain a real-time speech enhancement system. With 3.58M parameters and a low latency, our system is designed for real-time streaming with a lightweight architecture. Experiments show that our system improves over the first stage in terms of NISQA-MOS metric. Finally, through an ablation study, we show the importance of noisy conditioning in our system. We participated in 2025 Urgent Challenge with our model and later made further improvements.
Authors: Fengxiang Wang, Mingshuo Chen, Xuming He, YiFan Zhang, Feng Liu, Zijie Guo, Zhenghao Hu, Jiong Wang, Jingyi Xu, Zhangrui Li, Fenghua Ling, Ben Fei, Weijia Li, Long Lan, Wenjing Yang, Wenlong Zhang, Lei Bai
Abstract: Existing benchmarks for Earth science multimodal learning exhibit critical limitations in systematic coverage of geosystem components and cross-sphere interactions, often constrained to isolated subsystems (only in Human-activities sphere or atmosphere) with limited evaluation dimensions (less than 16 tasks). To address these gaps, we introduce OmniEarth-Bench, the first comprehensive multimodal benchmark spanning all six Earth science spheres (atmosphere, lithosphere, Oceansphere, cryosphere, biosphere and Human-activities sphere) and cross-spheres with one hundred expert-curated evaluation dimensions. Leveraging observational data from satellite sensors and in-situ measurements, OmniEarth-Bench integrates 29,779 annotations across four tiers: perception, general reasoning, scientific knowledge reasoning and chain-of-thought (CoT) reasoning. This involves the efforts of 2-5 experts per sphere to establish authoritative evaluation dimensions and curate relevant observational datasets, 40 crowd-sourcing annotators to assist experts for annotations, and finally, OmniEarth-Bench is validated via hybrid expert-crowd workflows to reduce label ambiguity. Experiments on 9 state-of-the-art MLLMs reveal that even the most advanced models struggle with our benchmarks, where none of them reach 35\% accuracy. Especially, in some cross-spheres tasks, the performance of leading models like GPT-4o drops to 0.0\%. OmniEarth-Bench sets a new standard for geosystem-aware AI, advancing both scientific discovery and practical applications in environmental monitoring and disaster prediction. The dataset, source code, and trained models were released.
Authors: Hayden Moore, Sirui Qi, Ninad Hogade, Dejan Milojicic, Cullen Bash, Sudeep Pasricha
Abstract: In recent years, Large Language Models (LLM) such as ChatGPT, CoPilot, and Gemini have been widely adopted in different areas. As the use of LLMs continues to grow, many efforts have focused on reducing the massive training overheads of these models. But it is the environmental impact of handling user requests to LLMs that is increasingly becoming a concern. Recent studies estimate that the costs of operating LLMs in their inference phase can exceed training costs by 25x per year. As LLMs are queried incessantly, the cumulative carbon footprint for the operational phase has been shown to far exceed the footprint during the training phase. Further, estimates indicate that 500 ml of fresh water is expended for every 20-50 requests to LLMs during inference. To address these important sustainability issues with LLMs, we propose a novel framework called SLIT to co-optimize LLM quality of service (time-to-first token), carbon emissions, water usage, and energy costs. The framework utilizes a machine learning (ML) based metaheuristic to enhance the sustainability of LLM hosting across geo-distributed cloud datacenters. Such a framework will become increasingly vital as LLMs proliferate.
Authors: Marc Jourdan, Gizem Y\"uce, Nicolas Flammarion
Abstract: Recent advances in language modeling have underscored the role of preference feedback in enhancing model performance. This paper investigates the conditions under which preference feedback improves parameter estimation in classes of continuous parametric distributions. In our framework, the learner observes pairs of samples from an unknown distribution along with their relative preferences depending on the same unknown parameter. We show that preference-based M-estimators achieve a better asymptotic variance than sample-only M-estimators, further improved by deterministic preferences. Leveraging the hard constraints revealed by deterministic preferences, we propose an estimator achieving an estimation error scaling of $\mathcal{O}(1/n)$ -- a significant improvement over the $\Theta(1/\sqrt{n})$ rate attainable with samples alone. Next, we establish a lower bound that matches this accelerated rate; up to dimension and problem-dependent constants. While the assumptions underpinning our analysis are restrictive, they are satisfied by notable cases such as Gaussian or Laplace distributions for preferences based on the log-probability reward.
Authors: Xu Chu, Xinrong Chen, Guanyu Wang, Zhijie Tan, Kui Huang, Wenyu Lv, Tong Mo, Weiping Li
Abstract: Inference time scaling drives extended reasoning to enhance the performance of Vision-Language Models (VLMs), thus forming powerful Vision-Language Reasoning Models (VLRMs). However, long reasoning dilutes visual tokens, causing visual information to receive less attention and may trigger hallucinations. Although introducing text-only reflection processes shows promise in language models, we demonstrate that it is insufficient to suppress hallucinations in VLMs. To address this issue, we introduce Qwen-LookAgain (Qwen-LA), a novel VLRM designed to mitigate hallucinations by incorporating a vision-text reflection process that guides the model to re-attention visual information during reasoning. We first propose a reinforcement learning method Balanced Reflective Policy Optimization (BRPO), which guides the model to decide when to generate vision-text reflection on its own and balance the number and length of reflections. Then, we formally prove that VLRMs lose attention to visual tokens as reasoning progresses, and demonstrate that supplementing visual information during reflection enhances visual attention. Therefore, during training and inference, Visual Token COPY and Visual Token ROUTE are introduced to force the model to re-attention visual information at the visual level, addressing the limitations of text-only reflection. Experiments on multiple visual QA datasets and hallucination metrics indicate that Qwen-LA achieves leading accuracy performance while reducing hallucinations. Our code is available at: https://github.com/Liar406/Look_Again.
Authors: Benjamin Arnav, Pablo Bernabeu-P\'erez, Nathan Helm-Burger, Tim Kostolansky, Hannes Whittingham, Mary Phuong
Abstract: As AI models are deployed with increasing autonomy, it is important to ensure they do not take harmful actions unnoticed. As a potential mitigation, we investigate Chain-of-Thought (CoT) monitoring, wherein a weaker trusted monitor model continuously oversees the intermediate reasoning steps of a more powerful but untrusted model. We compare CoT monitoring to action-only monitoring, where only final outputs are reviewed, in a red-teaming setup where the untrusted model is instructed to pursue harmful side tasks while completing a coding problem. We find that CoT monitoring improves detection by up to 27 percentage points in scenarios where action-only monitoring fails to reliably identify sabotage. However, CoT traces can also contain misleading rationalizations that deceive the monitor, reducing performance in more obvious sabotage cases. To address this, we introduce a hybrid protocol that independently scores both reasoning and final outputs and combines them using a weighted average. This hybrid monitor consistently outperforms both CoT and action-only monitors across all tested models and tasks, with detection rates over four times higher than action-only monitoring for subtle deception scenarios.
Authors: Christian Schmidt, Heinrich Martin Overhoff
Abstract: In medical image segmentation, limited external validity remains a critical obstacle when models are deployed across unseen datasets, an issue particularly pronounced in the ultrasound image domain. Existing solutions-such as domain adaptation and GAN-based style transfer-while promising, often fall short in the medical domain where datasets are typically small and diverse. This paper presents a novel application of principal component analysis (PCA) to address this limitation. PCA preprocessing reduces noise and emphasizes essential features by retaining approximately 90\% of the dataset variance. We evaluate our approach across six diverse breast tumor ultrasound datasets comprising 3,983 B-mode images and corresponding expert tumor segmentation masks. For each dataset, a corresponding dimensionality reduced PCA-dataset is created and U-Net-based segmentation models are trained on each of the twelve datasets. Each model trained on an original dataset was inferenced on the remaining five out-of-domain original datasets (baseline results), while each model trained on a PCA dataset was inferenced on five out-of-domain PCA datasets. Our experimental results indicate that using PCA reconstructed datasets, instead of original images, improves the model's recall and Dice scores, particularly for model-dataset pairs where baseline performance was lowest, achieving statistically significant gains in recall (0.57 $\pm$ 0.07 vs. 0.70 $\pm$ 0.05, $p = 0.0004$) and Dice scores (0.50 $\pm$ 0.06 vs. 0.58 $\pm$ 0.06, $p = 0.03$). Our method reduced the decline in recall values due to external validation by $33\%$. These findings underscore the potential of PCA reconstruction as a safeguard to mitigate declines in segmentation performance, especially in challenging cases, with implications for enhancing external validity in real-world medical applications.
Authors: Xi Chen, Soham Jana, Christopher A. Metzler, Arian Maleki, Shirin Jalali
Abstract: Multilook coherent imaging is a widely used technique in applications such as digital holography, ultrasound imaging, and synthetic aperture radar. A central challenge in these systems is the presence of multiplicative noise, commonly known as speckle, which degrades image quality. Despite the widespread use of coherent imaging systems, their theoretical foundations remain relatively underexplored. In this paper, we study both the theoretical and algorithmic aspects of likelihood-based approaches for multilook coherent imaging, providing a rigorous framework for analysis and method development. Our theoretical contributions include establishing the first theoretical upper bound on the Mean Squared Error (MSE) of the maximum likelihood estimator under the deep image prior hypothesis. Our results capture the dependence of MSE on the number of parameters in the deep image prior, the number of looks, the signal dimension, and the number of measurements per look. On the algorithmic side, we employ projected gradient descent (PGD) as an efficient method for computing the maximum likelihood solution. Furthermore, we introduce two key ideas to enhance the practical performance of PGD. First, we incorporate the Newton-Schulz algorithm to compute matrix inverses within the PGD iterations, significantly reducing computational complexity. Second, we develop a bagging strategy to mitigate projection errors introduced during PGD updates. We demonstrate that combining these techniques with PGD yields state-of-the-art performance. Our code is available at https://github.com/Computational-Imaging-RU/Bagged-DIP-Speckle.
URLs: https://github.com/Computational-Imaging-RU/Bagged-DIP-Speckle.
Authors: Chenhao Zheng, Jieyu Zhang, Mohammadreza Salehi, Ziqi Gao, Vishnu Iyengar, Norimasa Kobori, Quan Kong, Ranjay Krishna
Abstract: Effective video tokenization is critical for scaling transformer models for long videos. Current approaches tokenize videos using space-time patches, leading to excessive tokens and computational inefficiencies. The best token reduction strategies degrade performance and barely reduce the number of tokens when the camera moves. We introduce grounded video tokenization, a paradigm that organizes tokens based on panoptic sub-object trajectories rather than fixed patches. Our method aligns with fundamental perceptual principles, ensuring that tokenization reflects scene complexity rather than video duration. We propose TrajViT, a video encoder that extracts object trajectories and converts them into semantically meaningful tokens, significantly reducing redundancy while maintaining temporal coherence. Trained with contrastive learning, TrajViT significantly outperforms space-time ViT (ViT3D) across multiple video understanding benchmarks, e.g., TrajViT outperforms ViT3D by a large margin of 6% top-5 recall in average at video-text retrieval task with 10x token deduction. We also show TrajViT as a stronger model than ViT3D for being the video encoder for modern VideoLLM, obtaining an average of 5.2% performance improvement across 6 VideoQA benchmarks while having 4x faster training time and 18x less inference FLOPs. TrajViT is the first efficient encoder to consistently outperform ViT3D across diverse video analysis tasks, making it a robust and scalable solution.
Authors: Neta Glazer, David Chernin, Idan Achituve, Sharon Gannot, Ethan Fetaya
Abstract: Recent advancements in Text-to-Speech (TTS) models, particularly in voice cloning, have intensified the demand for adaptable and efficient deepfake detection methods. As TTS systems continue to evolve, detection models must be able to efficiently adapt to previously unseen generation models with minimal data. This paper introduces ADD-GP, a few-shot adaptive framework based on a Gaussian Process (GP) classifier for Audio Deepfake Detection (ADD). We show how the combination of a powerful deep embedding model with the Gaussian processes flexibility can achieve strong performance and adaptability. Additionally, we show this approach can also be used for personalized detection, with greater robustness to new TTS models and one-shot adaptability. To support our evaluation, a benchmark dataset is constructed for this task using new state-of-the-art voice cloning models.
Authors: Jiayuan Ye, Vitaly Feldman, Kunal Talwar
Abstract: We study the fundamental problem of estimating an unknown discrete distribution $p$ over $d$ symbols, given $n$ i.i.d. samples from the distribution. We are interested in minimizing the KL divergence between the true distribution and the algorithm's estimate. We first construct minimax optimal private estimators. Minimax optimality however fails to shed light on an algorithm's performance on individual (non-worst-case) instances $p$ and simple minimax-optimal DP estimators can have poor empirical performance on real distributions. We then study this problem from an instance-optimality viewpoint, where the algorithm's error on $p$ is compared to the minimum achievable estimation error over a small local neighborhood of $p$. Under natural notions of local neighborhood, we propose algorithms that achieve instance-optimality up to constant factors, with and without a differential privacy constraint. Our upper bounds rely on (private) variants of the Good-Turing estimator. Our lower bounds use additive local neighborhoods that more precisely captures the hardness of distribution estimation in KL divergence, compared to ones considered in prior works.
Authors: Zijun Yao, Yantao Liu, Yanxu Chen, Jianhui Chen, Junfeng Fang, Lei Hou, Juanzi Li, Tat-Seng Chua
Abstract: Recently evolved large reasoning models (LRMs) show powerful performance in solving complex tasks with long chain-of-thought (CoT) reasoning capability. As these LRMs are mostly developed by post-training on formal reasoning tasks, whether they generalize the reasoning capability to help reduce hallucination in fact-seeking tasks remains unclear and debated. For instance, DeepSeek-R1 reports increased performance on SimpleQA, a fact-seeking benchmark, while OpenAI-o3 observes even severer hallucination. This discrepancy naturally raises the following research question: Are reasoning models more prone to hallucination? This paper addresses the question from three perspectives. (1) We first conduct a holistic evaluation for the hallucination in LRMs. Our analysis reveals that LRMs undergo a full post-training pipeline with cold start supervised fine-tuning (SFT) and verifiable reward RL generally alleviate their hallucination. In contrast, both distillation alone and RL training without cold start fine-tuning introduce more nuanced hallucinations. (2) To explore why different post-training pipelines alters the impact on hallucination in LRMs, we conduct behavior analysis. We characterize two critical cognitive behaviors that directly affect the factuality of a LRM: Flaw Repetition, where the surface-level reasoning attempts repeatedly follow the same underlying flawed logic, and Think-Answer Mismatch, where the final answer fails to faithfully match the previous CoT process. (3) Further, we investigate the mechanism behind the hallucination of LRMs from the perspective of model uncertainty. We find that increased hallucination of LRMs is usually associated with the misalignment between model uncertainty and factual accuracy. Our work provides an initial understanding of the hallucination in LRMs.
Authors: Yuehaw Khoo, Mathias Oster, Yifan Peng
Abstract: Diffusion models have emerged as a powerful framework in generative modeling, typically relying on optimizing neural networks to estimate the score function via forward SDE simulations. In this work, we propose an alternative method that is both optimization-free and forward SDE-free. By expanding the score function in a sparse set of eigenbasis of the backward Kolmogorov operator associated with the diffusion process, we reformulate score estimation as the solution to a linear system, avoiding iterative optimization and time-dependent sample generation. We analyze the approximation error using perturbation theory and demonstrate the effectiveness of our method on high-dimensional Boltzmann distributions and real-world datasets.
Authors: Hongxiang Zhang, Hao Chen, Tianyi Zhang, Muhao Chen
Abstract: Recent decoding methods improve the factuality of large language models~(LLMs) by refining how the next token is selected during generation. These methods typically operate at the token level, leveraging internal representations to suppress superficial patterns. Nevertheless, LLMs remain prone to hallucinations, especially over longer contexts. In this paper, we propose Active Layer-Contrastive Decoding (ActLCD), a novel decoding strategy that actively decides when to apply contrasting layers during generation. By casting decoding as a sequential decision-making problem, ActLCD employs a reinforcement learning policy guided by a reward-aware classifier to optimize factuality beyond the token level. Our experiments demonstrate that ActLCD surpasses state-of-the-art methods across five benchmarks, showcasing its effectiveness in mitigating hallucinations in diverse generation scenarios.
Authors: Haim Kaplan, Yishay Mansour, Kobbi Nissim, Uri Stemmer
Abstract: We introduce a new Bayesian perspective on the concept of data reconstruction, and leverage this viewpoint to propose a new security definition that, in certain settings, provably prevents reconstruction attacks. We use our paradigm to shed new light on one of the most notorious attacks in the privacy and memorization literature - fingerprinting code attacks (FPC). We argue that these attacks are really a form of membership inference attacks, rather than reconstruction attacks. Furthermore, we show that if the goal is solely to prevent reconstruction (but not membership inference), then in some cases the impossibility results derived from FPC no longer apply.
Authors: Luke McDermott, Robert W. Heath Jr., Rahul Parhi
Abstract: Transformer-based large language models suffer from quadratic complexity at inference on long sequences. Linear attention methods are efficient alternatives, however, they fail to provide an accurate approximation of softmax attention. By additionally incorporating sliding window attention into each linear attention head, this gap can be closed for short context-length tasks. Unfortunately, these approaches cannot recall important information from long contexts due to "memory collisions". In this paper , we propose LoLA: Low-rank Linear Attention with sparse caching. LoLA separately stores additional key-value pairs that would otherwise interfere with past associative memories. Moreover, LoLA further closes the gap between linear attention models and transformers by distributing past key-value pairs into three forms of memory: (i) recent pairs in a local sliding window; (ii) difficult-to-memorize pairs in a sparse, global cache; and (iii) generic pairs in the recurrent hidden state of linear attention. As an inference-only strategy, LoLA enables pass-key retrieval on up to 8K context lengths on needle-in-a-haystack tasks from RULER. It boosts the accuracy of the base subquadratic model from 0.6% to 97.4% at 4K context lengths, with a 4.6x smaller cache than that of Llama-3.1 8B. LoLA demonstrates strong performance on zero-shot commonsense reasoning tasks among 1B and 8B parameter subquadratic models. Finally, LoLA is an extremely lightweight approach: Nearly all of our results can be reproduced on a single consumer GPU.
Authors: Manish Shetty, Naman Jain, Jinjian Liu, Vijay Kethanaboyina, Koushik Sen, Ion Stoica
Abstract: Developing high-performance software is a complex task that requires specialized expertise. We introduce GSO, a benchmark for evaluating language models' capabilities in developing high-performance software. We develop an automated pipeline that generates and executes performance tests to analyze repository commit histories to identify 102 challenging optimization tasks across 10 codebases, spanning diverse domains and programming languages. An agent is provided with a codebase and performance test as a precise specification, and tasked to improve the runtime efficiency, which is measured against the expert developer optimization. Our quantitative evaluation reveals that leading SWE-Agents struggle significantly, achieving less than 5% success rate, with limited improvements even with inference-time scaling. Our qualitative analysis identifies key failure modes, including difficulties with low-level languages, practicing lazy optimization strategies, and challenges in accurately localizing bottlenecks. We release the code and artifacts of our benchmark along with agent trajectories to enable future research.
Authors: Jingyun Yang, Isabella Huang, Brandon Vu, Max Bajracharya, Rika Antonova, Jeannette Bohg
Abstract: Learned visuomotor policies are capable of performing increasingly complex manipulation tasks. However, most of these policies are trained on data collected from limited robot positions and camera viewpoints. This leads to poor generalization to novel robot positions, which limits the use of these policies on mobile platforms, especially for precise tasks like pressing buttons or turning faucets. In this work, we formulate the policy mobilization problem: find a mobile robot base pose in a novel environment that is in distribution with respect to a manipulation policy trained on a limited set of camera viewpoints. Compared to retraining the policy itself to be more robust to unseen robot base pose initializations, policy mobilization decouples navigation from manipulation and thus does not require additional demonstrations. Crucially, this problem formulation complements existing efforts to improve manipulation policy robustness to novel viewpoints and remains compatible with them. To study policy mobilization, we introduce the Mobi-$\pi$ framework, which includes: (1) metrics that quantify the difficulty of mobilizing a given policy, (2) a suite of simulated mobile manipulation tasks based on RoboCasa to evaluate policy mobilization, (3) visualization tools for analysis, and (4) several baseline methods. We also propose a novel approach that bridges navigation and manipulation by optimizing the robot's base pose to align with an in-distribution base pose for a learned policy. Our approach utilizes 3D Gaussian Splatting for novel view synthesis, a score function to evaluate pose suitability, and sampling-based optimization to identify optimal robot poses. We show that our approach outperforms baselines in both simulation and real-world environments, demonstrating its effectiveness for policy mobilization.
Authors: Dionysis Christopoulos, Sotiris Spanos, Eirini Baltzi, Valsamis Ntouskos, Konstantinos Karantzalos
Abstract: We introduce SLIMP (Skin Lesion Image-Metadata Pre-training) for learning rich representations of skin lesions through a novel nested contrastive learning approach that captures complex relationships between images and metadata. Melanoma detection and skin lesion classification based solely on images, pose significant challenges due to large variations in imaging conditions (lighting, color, resolution, distance, etc.) and lack of clinical and phenotypical context. Clinicians typically follow a holistic approach for assessing the risk level of the patient and for deciding which lesions may be malignant and need to be excised, by considering the patient's medical history as well as the appearance of other lesions of the patient. Inspired by this, SLIMP combines the appearance and the metadata of individual skin lesions with patient-level metadata relating to their medical record and other clinically relevant information. By fully exploiting all available data modalities throughout the learning process, the proposed pre-training strategy improves performance compared to other pre-training strategies on downstream skin lesions classification tasks highlighting the learned representations quality.
Authors: Zexi Liu, Jingyi Chai, Xinyu Zhu, Shuo Tang, Rui Ye, Bo Zhang, Lei Bai, Siheng Chen
Abstract: The emergence of large language model (LLM)-based agents has significantly advanced the development of autonomous machine learning (ML) engineering. However, most existing approaches rely heavily on manual prompt engineering, failing to adapt and optimize based on diverse experimental experiences. Focusing on this, for the first time, we explore the paradigm of learning-based agentic ML, where an LLM agent learns through interactive experimentation on ML tasks using online reinforcement learning (RL). To realize this, we propose a novel agentic ML training framework with three key components: (1) exploration-enriched fine-tuning, which enables LLM agents to generate diverse actions for enhanced RL exploration; (2) step-wise RL, which enables training on a single action step, accelerating experience collection and improving training efficiency; (3) an agentic ML-specific reward module, which unifies varied ML feedback signals into consistent rewards for RL optimization. Leveraging this framework, we train ML-Agent, driven by a 7B-sized Qwen-2.5 LLM for autonomous ML. Remarkably, despite being trained on merely 9 ML tasks, our 7B-sized ML-Agent outperforms the 671B-sized DeepSeek-R1 agent. Furthermore, it achieves continuous performance improvements and demonstrates exceptional cross-task generalization capabilities.
Authors: Wei Shen, Ruichuan Huang, Minhui Huang, Cong Shen, Jiawei Zhang
Abstract: The majority of parameters in neural networks are naturally represented as matrices. However, most commonly used optimizers treat these matrix parameters as flattened vectors during optimization, potentially overlooking their inherent structural properties. Recently, an optimizer called Muon has been proposed, specifically designed to optimize matrix-structured parameters. Extensive empirical evidence shows that Muon can significantly outperform traditional optimizers when training neural networks. Nonetheless, the theoretical understanding of Muon's convergence behavior and the reasons behind its superior performance remain limited. In this work, we present a comprehensive convergence rate analysis of Muon and its comparison with Gradient Descent (GD). We further characterize the conditions under which Muon can outperform GD. Our theoretical results reveal that Muon can benefit from the low-rank and approximate blockwise diagonal structure of Hessian matrices -- phenomena widely observed in practical neural network training. Our experimental results support and corroborate the theoretical findings.
Authors: Hao Dong, Moru Liu, Jian Liang, Eleni Chatzi, Olga Fink
Abstract: Vision-Language Models (VLMs) have demonstrated strong capabilities in aligning visual and textual modalities, enabling a wide range of applications in multimodal understanding and generation. While they excel in zero-shot and transfer learning scenarios, VLMs remain susceptible to misclassification, often yielding confident yet incorrect predictions. This limitation poses a significant risk in safety-critical domains, where erroneous predictions can lead to severe consequences. In this work, we introduce TrustVLM, a training-free framework designed to address the critical challenge of estimating when VLM's predictions can be trusted. Motivated by the observed modality gap in VLMs and the insight that certain concepts are more distinctly represented in the image embedding space, we propose a novel confidence-scoring function that leverages this space to improve misclassification detection. We rigorously evaluate our approach across 17 diverse datasets, employing 4 architectures and 2 VLMs, and demonstrate state-of-the-art performance, with improvements of up to 51.87% in AURC, 9.14% in AUROC, and 32.42% in FPR95 compared to existing baselines. By improving the reliability of the model without requiring retraining, TrustVLM paves the way for safer deployment of VLMs in real-world applications. The code will be available at https://github.com/EPFL-IMOS/TrustVLM.
Authors: Diankun Wu, Fangfu Liu, Yi-Hsin Hung, Yueqi Duan
Abstract: Recent advancements in Multimodal Large Language Models (MLLMs) have significantly enhanced performance on 2D visual tasks. However, improving their spatial intelligence remains a challenge. Existing 3D MLLMs always rely on additional 3D or 2.5D data to incorporate spatial awareness, restricting their utility in scenarios with only 2D inputs, such as images or videos. In this paper, we present Spatial-MLLM, a novel framework for visual-based spatial reasoning from purely 2D observations. Unlike conventional video MLLMs which rely on CLIP-based visual encoders optimized for semantic understanding, our key insight is to unleash the strong structure prior from the feed-forward visual geometry foundation model. Specifically, we propose a dual-encoder architecture: a pretrained 2D visual encoder to extract semantic features, and a spatial encoder-initialized from the backbone of the visual geometry model-to extract 3D structure features. A connector then integrates both features into unified visual tokens for enhanced spatial understanding. Furthermore, we propose a space-aware frame sampling strategy at inference time, which selects the spatially informative frames of a video sequence, ensuring that even under limited token length, the model focuses on frames critical for spatial reasoning. Beyond architecture improvements, we construct the Spatial-MLLM-120k dataset and train the model on it using supervised fine-tuning and GRPO. Extensive experiments on various real-world datasets demonstrate that our spatial-MLLM achieves state-of-the-art performance in a wide range of visual-based spatial understanding and reasoning tasks. Project page: https://diankun-wu.github.io/Spatial-MLLM/.
Authors: Heekyung Lee, Jiaxin Ge, Tsung-Han Wu, Minwoo Kang, Trevor Darrell, David M. Chan
Abstract: Rebus puzzles, visual riddles that encode language through imagery, spatial arrangement, and symbolic substitution, pose a unique challenge to current vision-language models (VLMs). Unlike traditional image captioning or question answering tasks, rebus solving requires multi-modal abstraction, symbolic reasoning, and a grasp of cultural, phonetic and linguistic puns. In this paper, we investigate the capacity of contemporary VLMs to interpret and solve rebus puzzles by constructing a hand-generated and annotated benchmark of diverse English-language rebus puzzles, ranging from simple pictographic substitutions to spatially-dependent cues ("head" over "heels"). We analyze how different VLMs perform, and our findings reveal that while VLMs exhibit some surprising capabilities in decoding simple visual clues, they struggle significantly with tasks requiring abstract reasoning, lateral thinking, and understanding visual metaphors.
Authors: Wentao Zhang, Woojeong Kim, Yuntian Deng
Abstract: Conversational agents powered by large language models (LLMs) are rapidly becoming integral to our daily interactions, generating unprecedented amounts of conversational data. Such datasets offer a powerful lens into societal interests, trending topics, and collective concerns. Yet, existing approaches typically treat these interactions as independent and miss critical insights that could emerge from aggregating and reasoning across large-scale conversation logs. In this paper, we introduce Aggregative Question Answering, a novel task requiring models to reason explicitly over thousands of user-chatbot interactions to answer aggregative queries, such as identifying emerging concerns among specific demographics. To enable research in this direction, we construct a benchmark, WildChat-AQA, comprising 6,027 aggregative questions derived from 182,330 real-world chatbot conversations. Experiments show that existing methods either struggle to reason effectively or incur prohibitive computational costs, underscoring the need for new approaches capable of extracting collective insights from large-scale conversational data.
Authors: Jacob Beck, Risto Vuorio, Evan Zheran Liu, Zheng Xiong, Luisa Zintgraf, Chelsea Finn, Shimon Whiteson
Abstract: While deep reinforcement learning (RL) has fueled multiple high-profile successes in machine learning, it is held back from more widespread adoption by its often poor data efficiency and the limited generality of the policies it produces. A promising approach for alleviating these limitations is to cast the development of better RL algorithms as a machine learning problem itself in a process called meta-RL. Meta-RL is most commonly studied in a problem setting where, given a distribution of tasks, the goal is to learn a policy that is capable of adapting to any new task from the task distribution with as little data as possible. In this survey, we describe the meta-RL problem setting in detail as well as its major variations. We discuss how, at a high level, meta-RL research can be clustered based on the presence of a task distribution and the learning budget available for each individual task. Using these clusters, we then survey meta-RL algorithms and applications. We conclude by presenting the open problems on the path to making meta-RL part of the standard toolbox for a deep RL practitioner.
Authors: S. Abdurakhmanova, Y. SarcheshmehPour, A. Jung
Abstract: We present a model-agnostic federated learning method that mirrors the operation of a smart power grid: diverse local models, like energy prosumers, train independently on their own data while exchanging lightweight signals to coordinate with statistically similar peers. This coordination is governed by a graph-based regularizer that encourages connected models to produce similar predictions on a shared, public unlabeled dataset. The resulting method is a flexible instance of regularized empirical risk minimization and supports a wide variety of local models - both parametric and non-parametric - provided they can be trained via regularized loss minimization. Such training is readily supported by standard ML libraries including scikit-learn, Keras, and PyTorch.
Authors: David Liu, Jackie Baek, Tina Eliassi-Rad
Abstract: We study the fairness of dimensionality reduction methods for recommendations. We focus on the fundamental method of principal component analysis (PCA), which identifies latent components and produces a low-rank approximation via the leading components while discarding the trailing components. Prior works have defined notions of "fair PCA"; however, these definitions do not answer the following question: why is PCA unfair? We identify two underlying popularity mechanisms that induce item unfairness in PCA. The first negatively impacts less popular items because less popular items rely on trailing latent components to recover their values. The second negatively impacts highly popular items, since the leading PCA components specialize in individual popular items instead of capturing similarities between items. To address these issues, we develop a polynomial-time algorithm, Item-Weighted PCA, that flexibly up-weights less popular items when optimizing for leading principal components. We theoretically show that PCA, in all cases, and Normalized PCA, in cases of block-diagonal matrices, are instances of Item-Weighted PCA. We empirically show that there exist datasets for which Item-Weighted PCA yields the optimal solution while the baselines do not. In contrast to past dimensionality reduction re-weighting techniques, Item-Weighted PCA solves a convex optimization problem and enforces a hard rank constraint. Our evaluations on real-world datasets show that Item-Weighted PCA not only mitigates both unfairness mechanisms, but also produces recommendations that outperform those of PCA baselines.
Authors: Xiangyu Zhou, Yao Qiang, Saleh Zare Zade, Prashant Khanduri, Dongxiao Zhu
Abstract: In-context learning (ICL) has emerged as a powerful paradigm leveraging LLMs for specific downstream tasks by utilizing labeled examples as demonstrations (demos) in the preconditioned prompts. Despite its promising performance, crafted adversarial attacks pose a notable threat to the robustness of LLMs. Existing attacks are either easy to detect, require a trigger in user input, or lack specificity towards ICL. To address these issues, this work introduces a novel transferable prompt injection attack against ICL, aiming to hijack LLMs to generate the target output or elicit harmful responses. In our threat model, the hacker acts as a model publisher who leverages a gradient-based prompt search method to learn and append imperceptible adversarial suffixes to the in-context demos via prompt injection. We also propose effective defense strategies using a few shots of clean demos, enhancing the robustness of LLMs during ICL. Extensive experimental results across various classification and jailbreak tasks demonstrate the effectiveness of the proposed attack and defense strategies. This work highlights the significant security vulnerabilities of LLMs during ICL and underscores the need for further in-depth studies.
Authors: Ahmad Beirami, Alekh Agarwal, Jonathan Berant, Alexander D'Amour, Jacob Eisenstein, Chirag Nagpal, Ananda Theertha Suresh
Abstract: A simple and effective method for the inference-time alignment and scaling test-time compute of generative models is best-of-$n$ sampling, where $n$ samples are drawn from a reference policy, ranked based on a reward function, and the highest ranking one is selected. A commonly used analytical expression in the literature claims that the KL divergence between the best-of-$n$ policy and the reference policy is equal to $\log (n) - (n-1)/n.$ We disprove the validity of this claim, and show that it is an upper bound on the actual KL divergence. We also explore the tightness of this upper bound in different regimes, and propose a new estimator for the KL divergence and empirically show that it provides a tight approximation. We also show that the win rate of the best-of-$n$ policy against the reference policy is upper bounded by $n/(n+1)$ and derive bounds on the tightness of this characterization. We conclude with analyzing the tradeoffs between win rate and KL divergence of the best-of-$n$ alignment policy, which demonstrate that very good tradeoffs are achievable with $n < 1000$.
Authors: Matteo Gambella, Jary Pomponi, Simone Scardapane, Manuel Roveri
Abstract: Early Exit Neural Networks (EENNs) endow astandard Deep Neural Network (DNN) with Early Exit Classifiers (EECs), to provide predictions at intermediate points of the processing when enough confidence in classification is achieved. This leads to many benefits in terms of effectiveness and efficiency. Currently, the design of EENNs is carried out manually by experts, a complex and time-consuming task that requires accounting for many aspects, including the correct placement, the thresholding, and the computational overhead of the EECs. For this reason, the research is exploring the use of Neural Architecture Search (NAS) to automatize the design of EENNs. Currently, few comprehensive NAS solutions for EENNs have been proposed in the literature, and a fully automated, joint design strategy taking into consideration both the backbone and the EECs remains an open problem. To this end, this work presents Neural Architecture Search for Hardware Constrained Early Exit Neural Networks (NACHOS), the first NAS framework for the design of optimal EENNs satisfying constraints on the accuracy and the number of Multiply and Accumulate (MAC) operations performed by the EENNs at inference time. In particular, this provides the joint design of backbone and EECs to select a set of admissible (i.e., respecting the constraints) Pareto Optimal Solutions in terms of best tradeoff between the accuracy and number of MACs. The results show that the models designed by NACHOS are competitive with the state-of-the-art EENNs. Additionally, this work investigates the effectiveness of two novel regularization terms designed for the optimization of the auxiliary classifiers of the EENN
Authors: Keisuke Kawano, Takuro Kutsuna, Keisuke Sano
Abstract: Deep neural networks (DNNs) exhibit high performance in image recognition; however, the reasons for their strong generalization abilities remain unclear. A plausible hypothesis is that DNNs achieve robust and accurate predictions by identifying multiple pieces of evidence from images. Thus, to test this hypothesis, this study proposed minimal sufficient views (MSVs). MSVs is defined as a set of minimal regions within an input image that are sufficient to preserve the prediction of DNNs, thus representing the evidence discovered by the DNN. We empirically demonstrated a strong correlation between the number of MSVs (i.e., the number of pieces of evidence) and the generalization performance of the DNN models. Remarkably, this correlation was found to hold within a single DNN as well as between different DNNs, including convolutional and transformer models. This suggested that a DNN model that makes its prediction based on more evidence has a higher generalization performance. We proposed a metric based on MSVs for DNN model selection that did not require label information. Consequently, we empirically showed that the proposed metric was less dependent on the degree of overfitting, rendering it a more reliable indicator of model performance than existing metrics, such as average confidence.
Authors: Xiangyu Zhou, Yao Qiang, Saleh Zare Zade, Mohammad Amin Roshani, Prashant Khanduri, Douglas Zytko, Dongxiao Zhu
Abstract: The advent of Large Language Models (LLMs) has marked significant achievements in language processing and reasoning capabilities. Despite their advancements, LLMs face vulnerabilities to data poisoning attacks, where the adversary inserts backdoor triggers into training data to manipulate outputs. This work further identifies additional security risks in LLMs by designing a new data poisoning attack tailored to exploit the supervised fine-tuning (SFT) process. We propose a novel gradient-guided backdoor trigger learning (GBTL) algorithm to identify adversarial triggers efficiently, ensuring an evasion of detection by conventional defenses while maintaining content integrity. Through experimental validation across various language model tasks, including sentiment analysis, domain generation, and question answering, our poisoning strategy demonstrates a high success rate in compromising various LLMs' outputs. We further propose two defense strategies against data poisoning attacks, including in-context learning (ICL) and continuous learning (CL), which effectively rectify the behavior of LLMs and significantly reduce the decline in performance. Our work highlights the significant security risks present during SFT of LLMs and the necessity of safeguarding LLMs against data poisoning attacks.
Authors: Tianyu Chen, Haoyi Zhou, Ying Li, Hao Wang, Chonghan Gao, Rongye Shi, Shanghang Zhang, Jianxin Li
Abstract: Foundation models have revolutionized language modeling, while whether this success is replicated in scientific computing remains unexplored. We present OmniArch, the first prototype aiming at solving multi-scale and multi-physics scientific computing problems with physical alignment. We addressed all three challenges with one unified architecture. Its pre-training stage contains a Fourier Encoder-decoder fading out the disharmony across separated dimensions and a Transformer backbone integrating quantities through temporal dynamics, and the novel PDE-Aligner performs physics-informed fine-tuning under flexible conditions. As far as we know, we first conduct 1D-2D-3D united pre-training on the PDEBench, and it sets not only new performance benchmarks for 1D, 2D, and 3D PDEs but also demonstrates exceptional adaptability to new physics via in-context and zero-shot learning approaches, which supports realistic engineering applications and foresight physics discovery.
Authors: Daniele Angioni, Luca Demetrio, Maura Pintor, Luca Oneto, Davide Anguita, Battista Biggio, Fabio Roli
Abstract: Machine-learning models demand periodic updates to improve their average accuracy, exploiting novel architectures and additional data. However, a newly updated model may commit mistakes the previous model did not make. Such misclassifications are referred to as negative flips, experienced by users as a regression of performance. In this work, we show that this problem also affects robustness to adversarial examples, hindering the development of secure model update practices. In particular, when updating a model to improve its adversarial robustness, previously ineffective adversarial attacks on some inputs may become successful, causing a regression in the perceived security of the system. We propose a novel technique, named robustness-congruent adversarial training, to address this issue. It amounts to fine-tuning a model with adversarial training, while constraining it to retain higher robustness on the samples for which no adversarial example was found before the update. We show that our algorithm and, more generally, learning with non-regression constraints, provides a theoretically-grounded framework to train consistent estimators. Our experiments on robust models for computer vision confirm that both accuracy and robustness, even if improved after model update, can be affected by negative flips, and our robustness-congruent adversarial training can mitigate the problem, outperforming competing baseline methods.
Authors: Filip Szatkowski, Yaoyue Zheng, Fei Yang, Bart{\l}omiej Twardowski, Tomasz Trzci\'nski, Joost van de Weijer
Abstract: Continual learning is crucial for applying machine learning in challenging, dynamic, and often resource-constrained environments. However, catastrophic forgetting - overwriting previously learned knowledge when new information is acquired - remains a major challenge. In this work, we examine the intermediate representations in neural network layers during continual learning and find that such representations are less prone to forgetting, highlighting their potential to accelerate computation. Motivated by these findings, we propose to use auxiliary classifiers(ACs) to enhance performance and demonstrate that integrating ACs into various continual learning methods consistently improves accuracy across diverse evaluation settings, yielding an average 10% relative gain. We also leverage the ACs to reduce the average cost of the inference by 10-60% without compromising accuracy, enabling the model to return the predictions before computing all the layers. Our approach provides a scalable and efficient solution for continual learning.
Authors: Weilin Cai, Juyong Jiang, Le Qin, Junwei Cui, Sunghun Kim, Jiayi Huang
Abstract: Expert parallelism has emerged as a key strategy for distributing the computational workload of sparsely-gated mixture-of-experts (MoE) models across multiple devices, enabling the processing of increasingly large-scale models. However, the All-to-All communication inherent to expert parallelism poses a significant bottleneck, limiting the efficiency of MoE models. Although existing optimization methods partially mitigate this issue, they remain constrained by the sequential dependency between communication and computation operations. To address this challenge, we propose ScMoE, a novel shortcut-connected MoE architecture integrated with an overlapping parallelization strategy. ScMoE decouples communication from its conventional sequential ordering, enabling up to 100% overlap with computation. Compared to the prevalent top-2 MoE baseline, ScMoE achieves speedups of 1.49 times in training and 1.82 times in inference. Moreover, our experiments and analyses indicate that ScMoE not only achieves comparable but in some instances surpasses the model quality of existing approaches.
Authors: Kehua Feng, Keyan Ding, Hongzhi Tan, Kede Ma, Zhihua Wang, Shuangquan Guo, Yuzhou Cheng, Ge Sun, Guozhou Zheng, Qiang Zhang, Huajun Chen
Abstract: Reliable evaluation of large language models (LLMs) is impeded by two key challenges: objective metrics often fail to reflect human perception of natural language, and exhaustive human labeling is prohibitively expensive. Here, we propose a sample-efficient human evaluation method for LLMs based on the principle of MAximum Discrepancy (MAD) Competition. Our method automatically and adaptively selects a compact set of input instructions that maximize semantic discrepancy between pairs of LLM responses. Human evaluators then perform three-alternative forced choices on these paired responses, which are aggregated into a global ranking using Elo rating. We apply our approach to compare eight widely used LLMs across four tasks: scientific knowledge understanding, mathematical reasoning, creative and functional writing, and code generation and explanation. Experimental results show that our sample-efficient evaluation method recovers "gold-standard" model rankings with a handful of MAD-selected instructions, reveals respective strengths and weaknesses of each LLM, and offers nuanced insights to guide future LLM development. Code is available at https://github.com/weiji-Feng/MAD-Eval .
Authors: Tudor Cebere, Aur\'elien Bellet, Nicolas Papernot
Abstract: Machine learning models can be trained with formal privacy guarantees via differentially private optimizers such as DP-SGD. In this work, we focus on a threat model where the adversary has access only to the final model, with no visibility into intermediate updates. In the literature, this hidden state threat model exhibits a significant gap between the lower bound from empirical privacy auditing and the theoretical upper bound provided by privacy accounting. To challenge this gap, we propose to audit this threat model with adversaries that craft a gradient sequence designed to maximize the privacy loss of the final model without relying on intermediate updates. Our experiments show that this approach consistently outperforms previous attempts at auditing the hidden state model. Furthermore, our results advance the understanding of achievable privacy guarantees within this threat model. Specifically, when the crafted gradient is inserted at every optimization step, we show that concealing the intermediate model updates in DP-SGD does not enhance the privacy guarantees. The situation is more complex when the crafted gradient is not inserted at every step: our auditing lower bound matches the privacy upper bound only for an adversarially-chosen loss landscape and a sufficiently large batch size. This suggests that existing privacy upper bounds can be improved in certain regimes.
Authors: Shu Wei, Yanjie Li, Lina Yu, Weijun Li, Min Wu, Linjun Sun, Jufeng Han, Yan Pang
Abstract: The quest for analytical solutions to differential equations has traditionally been constrained by the need for extensive mathematical expertise. Machine learning methods like genetic algorithms have shown promise in this domain, but are hindered by significant computational time and the complexity of their derived solutions. This paper introduces SSDE (Symbolic Solver for Differential Equations), a novel reinforcement learning-based approach that derives symbolic closed-form solutions for various differential equations. Evaluations across a diverse set of ordinary and partial differential equations demonstrate that SSDE outperforms existing machine learning methods, delivering superior accuracy and efficiency in obtaining analytical solutions.
Authors: Oleksii Furman, Patryk Wielopolski, {\L}ukasz Lenkiewicz, Jerzy Stefanowski, Maciej Zi\k{e}ba
Abstract: The growing complexity of AI systems has intensified the need for transparency through Explainable AI (XAI). Counterfactual explanations (CFs) offer actionable "what-if" scenarios on three levels: Local CFs providing instance-specific insights, Global CFs addressing broader trends, and Group-wise CFs (GWCFs) striking a balance and revealing patterns within cohesive groups. Despite the availability of methods for each granularity level, the field lacks a unified method that integrates these complementary approaches. We address this limitation by proposing a gradient-based optimization method for differentiable models that generates Local, Global, and Group-wise Counterfactual Explanations in a unified manner. We especially enhance GWCF generation by combining instance grouping and counterfactual generation into a single efficient process, replacing traditional two-step methods. Moreover, to ensure trustworthiness, we innovatively introduce the integration of plausibility criteria into the GWCF domain, making explanations both valid and realistic. Our results demonstrate the method's effectiveness in balancing validity, proximity, and plausibility while optimizing group granularity, with practical utility validated through practical use cases.
Authors: Shuran Zheng, Xuan Qi, Rui Ray Chen, Yongchan Kwon, James Zou
Abstract: Data plays a central role in advancements in modern artificial intelligence, with high-quality data emerging as a key driver of model performance. This has prompted the development of principled and effective data curation methods in recent years. However, existing methods largely rely on heuristics, and whether they are truly effective remains unclear. For instance, standard evaluation methods that assess a trained model's performance on specific benchmarks may incentivize assigning high scores to data that merely resembles the test set. This issue exemplifies Goodhart's law: when a measure becomes a target, it ceases to be a good measure. To address this issue, we propose an information-theoretic framework for evaluating data curation methods. We define dataset quality in terms of its informativeness about the true model parameters, formalized using the Blackwell ordering of informativeness. Under this ordering, Blackwell's theorem ensures that more informative data yields optimal models with lower expected loss on the true underlying distribution. To measure informativeness, we show that the Blackwell order can be determined by the Shannon mutual information between the curated data and the test data. To estimate this mutual information, we introduce a novel method that trains Bayesian models on embedded datasets and computes mutual information from the posteriors of model parameters. Experiments on real-world data demonstrate that our mutual information-based evaluation assigns appropriately lower scores to data curation strategies that reduce dataset informativeness, while traditional test score-based evaluation methods may favor data curation strategies that overfit to the test set but compromise the training data's informativeness.
Authors: Yuansan Liu, Sudanthi Wijewickrema, Dongting Hu, Christofer Bester, Stephen O'Leary, James Bailey
Abstract: Recent innovations in diffusion probabilistic models have paved the way for significant progress in image, text and audio generation, leading to their applications in generative time series forecasting. However, leveraging such abilities to model highly stochastic time series data remains a challenge. In this paper, we propose a novel Stochastic Diffusion (StochDiff) model which learns data-driven prior knowledge at each time step by utilizing the representational power of the stochastic latent spaces to model the variability of the multivariate time series data. The learnt prior knowledge helps the model to capture complex temporal dynamics and the inherent uncertainty of the data. This improves its ability to model highly stochastic time series data. Through extensive experiments on real-world datasets, we demonstrate the effectiveness of our proposed model on stochastic time series forecasting. Additionally, we showcase an application of our model for real-world surgical guidance, highlighting its potential to benefit the medical community.
Authors: Zhen Xiang, Linzhi Zheng, Yanjie Li, Junyuan Hong, Qinbin Li, Han Xie, Jiawei Zhang, Zidi Xiong, Chulin Xie, Carl Yang, Dawn Song, Bo Li
Abstract: The rapid advancement of large language model (LLM) agents has raised new concerns regarding their safety and security. In this paper, we propose GuardAgent, the first guardrail agent to protect target agents by dynamically checking whether their actions satisfy given safety guard requests. Specifically, GuardAgent first analyzes the safety guard requests to generate a task plan, and then maps this plan into guardrail code for execution. By performing the code execution, GuardAgent can deterministically follow the safety guard request and safeguard target agents. In both steps, an LLM is utilized as the reasoning component, supplemented by in-context demonstrations retrieved from a memory module storing experiences from previous tasks. In addition, we propose two novel benchmarks: EICU-AC benchmark to assess the access control for healthcare agents and Mind2Web-SC benchmark to evaluate the safety policies for web agents. We show that GuardAgent effectively moderates the violation actions for different types of agents on these two benchmarks with over 98% and 83% guardrail accuracies, respectively. Project page: https://guardagent.github.io/
Authors: Niamh Belton, Aonghus Lawlor, Kathleen M. Curran
Abstract: The diagnostic accuracy and subjectivity of existing Knee Osteoarthritis (OA) ordinal grading systems has been a subject of on-going debate and concern. Existing automated solutions are trained to emulate these imperfect systems, whilst also being reliant on large annotated databases for fully-supervised training. This work proposes a three stage approach for automated continuous grading of knee OA that is built upon the principles of Anomaly Detection (AD); learning a robust representation of healthy knee X-rays and grading disease severity based on its distance to the centre of normality. In the first stage, SS-FewSOME is proposed, a self-supervised AD technique that learns the 'normal' representation, requiring only examples of healthy subjects and <3% of the labels that existing methods require. In the second stage, this model is used to pseudo label a subset of unlabelled data as 'normal' or 'anomalous', followed by denoising of pseudo labels with CLIP. The final stage involves retraining on labelled and pseudo labelled data using the proposed Dual Centre Representation Learning (DCRL) which learns the centres of two representation spaces; normal and anomalous. Disease severity is then graded based on the distance to the learned centres. The proposed methodology outperforms existing techniques by margins of up to 24% in terms of OA detection and the disease severity scores correlate with the Kellgren-Lawrence grading system at the same level as human expert performance. Code available at https://github.com/niamhbelton/SS-FewSOME_Disease_Severity_Knee_Osteoarthritis.
URLs: https://github.com/niamhbelton/SS-FewSOME_Disease_Severity_Knee_Osteoarthritis.
Authors: Yen-Ru Lai, Fu-Chieh Chang, Pei-Yuan Wu
Abstract: Offline reinforcement learning (RL) learns policies from a fixed dataset, but often requires large amounts of data. The challenge arises when labeled datasets are expensive, especially when rewards have to be provided by human labelers for large datasets. In contrast, unlabelled data tends to be less expensive. This situation highlights the importance of finding effective ways to use unlabelled data in offline RL, especially when labelled data is limited or expensive to obtain. In this paper, we present the algorithm to utilize the unlabeled data in the offline RL method with kernel function approximation and give the theoretical guarantee. We present various eigenvalue decay conditions of $\mathcal{H}_k$ which determine the complexity of the algorithm. In summary, our work provides a promising approach for exploiting the advantages offered by unlabeled data in offline RL, whilst maintaining theoretical assurances.
Authors: Arun Verma, Indrajit Saha, Makoto Yokoo, Bryan Kian Hsiang Low
Abstract: This paper considers a novel variant of the online fair division problem involving multiple agents in which a learner sequentially observes an indivisible item that has to be irrevocably allocated to one of the agents while satisfying a fairness and efficiency constraint. Existing algorithms assume a small number of items with a sufficiently large number of copies, which ensures a good utility estimation for all item-agent pairs from noisy bandit feedback. However, this assumption may not hold in many real-life applications, for example, an online platform that has a large number of users (items) who use the platform's service providers (agents) only a few times (a few copies of items), which makes it difficult to accurately estimate utilities for all item-agent pairs. To address this, we assume utility is an unknown function of item-agent features. We then propose algorithms that model online fair division as a contextual bandit problem, with sub-linear regret guarantees. Our experimental results further validate the effectiveness of the proposed algorithms.
Authors: Simran Kaur, Simon Park, Anirudh Goyal, Sanjeev Arora
Abstract: We introduce Instruct-SkillMix, an automated approach for creating diverse, high quality SFT data for instruction-following. The pipeline involves two stages, each leveraging an existing powerful LLM: (1) Skill extraction: uses the LLM to extract core "skills" for instruction-following by directly prompting the model. This is inspired by ``LLM metacognition'' of Didolkar et al. (2024); (2) Data generation: uses the powerful LLM to generate (instruction, response) data that exhibit a randomly chosen pair of these skills. Here, the use of random skill combinations promotes diversity and difficulty. The estimated cost of creating the dataset is under $600. Vanilla SFT (i.e., no PPO, DPO, or RL methods) on data generated from Instruct-SkillMix leads to strong gains on instruction following benchmarks such as AlpacaEval 2.0, MT-Bench, and WildBench. With just 4K examples, LLaMA-3-8B-Base achieves 42.76% length-controlled win rate on AlpacaEval 2.0, a level similar to frontier models like Claude 3 Opus and LLaMA-3.1-405B-Instruct. Ablation studies also suggest plausible reasons for why creating open instruction-tuning datasets via naive crowd-sourcing has proved difficult. In our dataset, adding 20% low quality answers (``shirkers'') causes a noticeable degradation in performance. The Instruct-SkillMix pipeline seems flexible and adaptable to other settings.
Authors: Shamsiiat Abdurakhmanova, Amirhossein Mohammadi, Yasmin SarcheshmehPour, Alexander Jung
Abstract: Many important machine learning applications involve networks of devices-such as wearables or smartphones-that generate local data and train personalized models. A key challenge is determining which peers are most beneficial for collaboration. We propose a simple and privacy-preserving method to select relevant collaborators by evaluating how much a model improves after a single gradient step using another devices data-without sharing raw data. This method naturally extends to non-parametric models by replacing the gradient step with a non-parametric generalization. Our approach enables model-agnostic, data-driven peer selection for personalized federated learning (PersFL).
Authors: Chuheng Zhang, Wei Shen, Li Zhao, Xuyun Zhang, Xiaolong Xu, Wanchun Dou, Jiang Biang
Abstract: While direct policy optimization methods exist, pioneering LLMs are fine-tuned with reinforcement learning from human feedback (RLHF) to generate better responses under the supervision of a reward model learned from preference data. One major challenge of RLHF is the inaccuracy of the intermediate reward model, especially in the tasks that requires complex reasoning for the reward model to score a response. We find that the reliability of the reward model varies across responses assigned with different rewards. This motivates us to filter the samples whose rewards may be unreliable to improve the signal-to-noise ratio during policy learning, resulting in Policy Filtration for Proximal Policy Optimization (PF-PPO). To choose a proper policy filtering strategy, we use the coefficient of determination (R2) between the rewards and actual scores on filtered samples as the metrics to help us find promising strategies since it measures how well the rewards filtered by PF-PPO indicate real performance. We provide extensive experiments to validate the effectiveness of PF-PPO in code generation and math reasoning tasks. In code generation, PF-PPO achieves the state-of-the-art performance of 7-billion-parameter models on HumanEval (+7.9%), MBPP (+0.7%), and LeetCode Contest (+10.0%) which is a more challenging benchmark created by us. In math reasoning, PF-PPO yields performance increase using different reward models and benchmarks (Ape210K and CMATH). Code is available on https://github.com/DtYXs/verl/tree/pf-ppo.
Authors: Dongyang Fan, Bettina Messmer, Nikita Doikov, Martin Jaggi
Abstract: On-device LLMs have gained increasing attention for their ability to enhance privacy and provide a personalized user experience. To facilitate private learning with scarce data, Federated Learning has become a standard approach. However, it faces challenges such as computational resource heterogeneity and data heterogeneity among end users. We propose CoMiGS ($\textbf{Co}$llaborative learning with a $\textbf{Mi}$xture of $\textbf{G}$eneralists and $\textbf{S}$pecialists), the first approach to address both challenges. A key innovation of our method is the bi-level optimization formulation of the Mixture-of-Experts learning objective, where the router is optimized using a separate validation set to ensure alignment with the target distribution. We solve our objective with alternating minimization, for which we provide a theoretical analysis. Our method shares generalist experts across users while localizing a varying number of specialist experts, thereby adapting to users' computational resources and preserving privacy. Through extensive experiments, we show CoMiGS effectively balances general and personalized knowledge for each token generation. We demonstrate that CoMiGS remains robust against overfitting-due to the generalists' regularizing effect-while adapting to local data through specialist expertise. We open source our codebase for collaborative LLMs.
Authors: Francesco Mori, Stefano Sarao Mannelli, Francesca Mignacco
Abstract: Artificial neural networks often struggle with catastrophic forgetting when learning multiple tasks sequentially, as training on new tasks degrades the performance on previously learned tasks. Recent theoretical work has addressed this issue by analysing learning curves in synthetic frameworks under predefined training protocols. However, these protocols relied on heuristics and lacked a solid theoretical foundation assessing their optimality. In this paper, we fill this gap by combining exact equations for training dynamics, derived using statistical physics techniques, with optimal control methods. We apply this approach to teacher-student models for continual learning and multi-task problems, obtaining a theory for task-selection protocols maximising performance while minimising forgetting. Our theoretical analysis offers non-trivial yet interpretable strategies for mitigating catastrophic forgetting, shedding light on how optimal learning protocols modulate established effects, such as the influence of task similarity on forgetting. Finally, we validate our theoretical findings with experiments on real-world data.
Authors: Shuoyuan Wang, Yixuan Li, Hongxin Wei
Abstract: Confidence calibration is critical for the safe deployment of machine learning models in the real world. However, such issue in vision-language models like CLIP, particularly after fine-tuning, has not been fully addressed. In this work, we demonstrate that existing prompt tuning methods usually lead to a trade-off of calibration between base and new classes: the cross-entropy loss in CoOp causes overconfidence in new classes by increasing textual label divergence, whereas the regularization of KgCoOp maintains the confidence level but results in underconfidence in base classes due to the improved accuracy. Inspired by the observations, we introduce Dynamic Outlier Regularization (DOR) to ensure the confidence calibration on both base and new classes after fine-tuning. In particular, we propose to minimize the feature deviation of novel textual labels (instead of base classes) sampled from a large vocabulary. In effect, DOR prevents the increase in textual divergence for new labels while easing restrictions on base classes. Extensive experiments demonstrate that DOR can enhance the calibration performance of current fine-tuning methods on base and new classes.
Authors: Aaditya Naik, Jason Liu, Claire Wang, Amish Sethi, Saikat Dutta, Mayur Naik, Eric Wong
Abstract: Neurosymbolic learning enables the integration of symbolic reasoning with deep learning but faces significant challenges in scaling to complex symbolic programs, large datasets, or both. We introduce DOLPHIN, a framework that tackles these challenges by supporting neurosymbolic programs in Python, executing complex symbolic reasoning on the CPU while vectorizing probabilistic computations and gradient propagation on the GPU. Across 13 benchmarks spanning tasks over text, image, and video data, with symbolic reasoning features like recursion and black-box functions, DOLPHIN converges to state-of-the-art accuracies on the more complex benchmarks while existing frameworks such as Scallop, ISED, and IndeCateR+ fail to converge within the time limit. On simpler benchmarks, DOLPHIN matches their performance, while achieving these results 1.71x to 62x faster than the baselines. Overall, DOLPHIN advances the scalability of neurosymbolic frameworks, achieving state-of-the-art efficiency and convergence on difficult benchmarks where existing frameworks struggle. The code is published at https://github.com/Dolphin-NeSy/Dolphin.
Authors: Ruifeng Ren, Zhicong Li, Yong Liu
Abstract: Transformers have become the backbone of modern Large Language Models (LLMs); however, their inference overhead grows linearly with the sequence length, posing challenges for modeling long sequences. In light of this, Mamba has attracted attention for maintaining a constant inference size, with empirical evidence demonstrating that it can match Transformer performance in sequence modeling while significantly reducing computational costs. However, an open question remains: can Mamba always bring savings while achieving performance comparable to Transformers? In this paper, we focus on analyzing the expressive ability of Mamba to perform our defined COPY operation and Chain of Thought (CoT) reasoning. First, inspired by the connection between Mamba and linear attention, we show that constant-sized Mamba may struggle to perform COPY operations while Transformers can handle them more easily. However, when the size of Mamba grows linearly with the input sequence length, it can accurately perform COPY, but in this case, Mamba no longer provides overhead savings. Based on this observation, we further analyze Mamba's ability to tackle CoT tasks, which can be described by the Dynamic Programming (DP) problems. Our findings suggest that to solve arbitrary DP problems, the total cost of Mamba is still comparable to standard Transformers. However, similar to efficient Transformers, when facing DP problems with favorable properties such as locality, Mamba can provide savings in overhead. Our experiments on the copy and CoT tasks further demonstrate Mamba's limitations compared to Transformers in learning these tasks.
Authors: Ann Huang, Satpreet H. Singh, Flavio Martinelli, Kanaka Rajan
Abstract: Task-trained recurrent neural networks (RNNs) are widely used in neuroscience and machine learning to model dynamical computations. To gain mechanistic insight into how neural systems solve tasks, prior work often reverse-engineers individual trained networks. However, different RNNs trained on the same task and achieving similar performance can exhibit strikingly different internal solutions-a phenomenon known as solution degeneracy. Here, we develop a unified framework to systematically quantify and control solution degeneracy across three levels: behavior, neural dynamics, and weight space. We apply this framework to 3,400 RNNs trained on four neuroscience-relevant tasks-flip-flop memory, sine wave generation, delayed discrimination, and path integration-while systematically varying task complexity, learning regime, network size, and regularization. We find that higher task complexity and stronger feature learning reduce degeneracy in neural dynamics but increase it in weight space, with mixed effects on behavior. In contrast, larger networks and structural regularization reduce degeneracy at all three levels. These findings empirically validate the Contravariance Principle and provide practical guidance for researchers aiming to tailor RNN solutions-whether to uncover shared neural mechanisms or to model individual variability observed in biological systems. This work provides a principled framework for quantifying and controlling solution degeneracy in task-trained RNNs, offering new tools for building more interpretable and biologically grounded models of neural computation.
Authors: Michael Munn, Susan Wei
Abstract: Recent advances in artificial intelligence have been fueled by the development of foundation models such as BERT, GPT, T5, and Vision Transformers. These models are first pretrained on vast and diverse datasets and then adapted to specific downstream tasks, often with significantly less data. However, the mechanisms behind the success of this ubiquitous pretrain-then-adapt paradigm remain underexplored, particularly the characteristics of pretraining checkpoints that enhance downstream adaptation. We introduce a Bayesian model selection criterion, called the downstream free energy, which quantifies a checkpoint's adaptability by measuring the concentration of nearby favorable parameters for the downstream task. We demonstrate that this Bayesian model selection criterion can be effectively implemented without access to the downstream data or prior knowledge of the downstream task. Furthermore, we provide empirical evidence that the criterion reliably correlates with improved finetuning performance, offering a principled approach to predicting model adaptability.
Authors: Mohammad Mozaffari, Amir Yazdanbakhsh, Maryam Mehri Dehnavi
Abstract: Conventional model compression techniques for LLMs address high memory consumption and slow inference challenges but typically require computationally expensive retraining to preserve accuracy. In contrast, one-shot compression methods eliminate retraining cost, but struggle to achieve accuracy comparable to dense models. This paper presents SLIM, a new one-shot compression framework that holistically integrates hardware-friendly quantization, sparsity, and low-rank approximation into a unified process. First, we formulate the quantization process using a probabilistic approach (SLIM-Quant) that enables us to apply uniform quantization. Then, we use an existing one-shot pruning method to apply semi-structured sparsity on top of the quantized weights. Finally, to compensate for the introduced aggregated quantization and sparsity error, we use a novel saliency function with unique invertible and additive features that enables us to mathematically compute the value of low-rank adapters. SLIM improves model accuracy by up to 5.66% (LLaMA-2-7B) for 2:4 sparsity with 4-bit weight quantization, outperforming prior methods. Models compressed with SLIM achieve up to 4.3x and 3.8x on Nvidia RTX3060 and A100 GPUs, respectively. Additionally, they achieve up to 0.23x end-to-end memory reduction in comparison to their dense counterparts. We also propose an optional PEFT recipe that further improves accuracy by up to 1.66% (LLaMA-2-13B) compared to SLIM without fine-tuning.
Authors: Hojoon Lee, Dongyoon Hwang, Donghu Kim, Hyunseung Kim, Jun Jet Tai, Kaushik Subramanian, Peter R. Wurman, Jaegul Choo, Peter Stone, Takuma Seno
Abstract: Recent advances in CV and NLP have been largely driven by scaling up the number of network parameters, despite traditional theories suggesting that larger networks are prone to overfitting. These large networks avoid overfitting by integrating components that induce a simplicity bias, guiding models toward simple and generalizable solutions. However, in deep RL, designing and scaling up networks have been less explored. Motivated by this opportunity, we present SimBa, an architecture designed to scale up parameters in deep RL by injecting a simplicity bias. SimBa consists of three components: (i) an observation normalization layer that standardizes inputs with running statistics, (ii) a residual feedforward block to provide a linear pathway from the input to output, and (iii) a layer normalization to control feature magnitudes. By scaling up parameters with SimBa, the sample efficiency of various deep RL algorithms-including off-policy, on-policy, and unsupervised methods-is consistently improved. Moreover, solely by integrating SimBa architecture into SAC, it matches or surpasses state-of-the-art deep RL methods with high computational efficiency across DMC, MyoSuite, and HumanoidBench. These results demonstrate SimBa's broad applicability and effectiveness across diverse RL algorithms and environments.
Authors: Wei Shen, Ruida Zhou, Jing Yang, Cong Shen
Abstract: Although transformers have demonstrated impressive capabilities for in-context learning (ICL) in practice, theoretical understanding of the underlying mechanism that allows transformers to perform ICL is still in its infancy. This work aims to theoretically study the training dynamics of transformers for in-context classification tasks. We demonstrate that, for in-context classification of Gaussian mixtures under certain assumptions, a single-layer transformer trained via gradient descent converges to a globally optimal model at a linear rate. We further quantify the impact of the training and testing prompt lengths on the ICL inference error of the trained transformer. We show that when the lengths of training and testing prompts are sufficiently large, the prediction of the trained transformer approaches the ground truth distribution of the labels. Experimental results corroborate the theoretical findings.
Authors: Minhua Lin, Zhiwei Zhang, Enyan Dai, Zongyu Wu, Yilong Wang, Xiang Zhang, Suhang Wang
Abstract: Graph Prompt Learning (GPL) has been introduced as a promising approach that uses prompts to adapt pre-trained GNN models to specific downstream tasks without requiring fine-tuning of the entire model. Despite the advantages of GPL, little attention has been given to its vulnerability to backdoor attacks, where an adversary can manipulate the model's behavior by embedding hidden triggers. Existing graph backdoor attacks rely on modifying model parameters during training, but this approach is impractical in GPL as GNN encoder parameters are frozen after pre-training. Moreover, downstream users may fine-tune their own task models on clean datasets, further complicating the attack. In this paper, we propose TGPA, a backdoor attack framework designed specifically for GPL. TGPA injects backdoors into graph prompts without modifying pre-trained GNN encoders and ensures high attack success rates and clean accuracy. To address the challenge of model fine-tuning by users, we introduce a finetuning-resistant poisoning approach that maintains the effectiveness of the backdoor even after downstream model adjustments. Extensive experiments on multiple datasets under various settings demonstrate the effectiveness of TGPA in compromising GPL models with fixed GNN encoders.
Authors: Peimeng Guan, Mark A. Davenport
Abstract: Inverse problems aim to reconstruct unseen data from corrupted or perturbed measurements. While most work focuses on improving reconstruction quality, generalization accuracy and robustness are equally important, especially for safety-critical applications. Model-based architectures (MBAs), such as loop unrolling methods, are considered more interpretable and achieve better reconstructions. Empirical evidence suggests that MBAs are more robust to perturbations than black-box solvers, but the accuracy-robustness tradeoff in MBAs remains underexplored. In this work, we propose a simple yet effective training scheme for MBAs, called SGD jittering, which injects noise iteration-wise during reconstruction. We theoretically demonstrate that SGD jittering not only generalizes better than the standard mean squared error training but is also more robust to average-case attacks. We validate SGD jittering using denoising toy examples, seismic deconvolution, and single-coil MRI reconstruction. Both SGD jittering and its SPGD extension yield cleaner reconstructions for out-of-distribution data and demonstrates enhanced robustness against adversarial attacks.
Authors: Bo Pan, Zhen Xiong, Guanchen Wu, Zheng Zhang, Yifei Zhang, Liang Zhao
Abstract: Graph representation learning has garnered significant attention due to its broad applications in various domains, such as recommendation systems and social network analysis. Despite advancements in graph learning methods, challenges still remain in explainability when graphs are associated with semantic features. In this paper, we present GraphNarrator, the first method designed to generate natural language explanations for Graph Neural Networks. GraphNarrator employs a generative language model that maps input-output pairs to explanations reflecting the model's decision-making process. To address the lack of ground truth explanations to train the model, we propose first generating pseudo-labels that capture the model's decisions from saliency-based explanations, then using Expert Iteration to iteratively train the pseudo-label generator based on training objectives on explanation quality. The high-quality pseudo-labels are finally utilized to train an end-to-end explanation generator model. Extensive experiments are conducted to demonstrate the effectiveness of GraphNarrator in producing faithful, concise, and human-preferred natural language explanations.
Authors: Anjiang Wei, Allen Nie, Thiago S. F. X. Teixeira, Rohan Yadav, Wonchan Lee, Ke Wang, Alex Aiken
Abstract: Modern scientific discovery increasingly relies on high-performance computing for complex modeling and simulation. A key challenge in improving parallel program performance is efficiently mapping tasks to processors and data to memory, a process dictated by intricate, low-level system code known as mappers. Developing high-performance mappers demands days of manual tuning, posing a significant barrier for domain scientists without systems expertise. We introduce a framework that automates mapper development with generative optimization, leveraging richer feedback beyond scalar performance metrics. Our approach features the Agent-System Interface, which includes a Domain-Specific Language (DSL) to abstract away the low-level complexity of system code and define a structured search space, as well as AutoGuide, a mechanism that interprets raw execution output into actionable feedback. Unlike traditional reinforcement learning methods such as OpenTuner, which rely solely on scalar feedback, our method finds superior mappers in far fewer iterations. With just 10 iterations, it outperforms OpenTuner even after 1000 iterations, achieving 3.8X faster performance. Our approach finds mappers that surpass expert-written mappers by up to 1.34X speedup across nine benchmarks while reducing tuning time from days to minutes.
Authors: Kento Nishi, Maya Okawa, Rahul Ramesh, Mikail Khona, Hidenori Tanaka, Ekdeep Singh Lubana
Abstract: Knowledge Editing (KE) algorithms alter models' weights to perform targeted updates to incorrect, outdated, or otherwise unwanted factual associations. However, recent work has shown that applying KE can adversely affect models' broader factual recall accuracy and diminish their reasoning abilities. Although these studies give insights into the potential harms of KE algorithms, e.g., performance evaluations on benchmarks, little is understood about why such destructive failures occur. Motivated by this, we define a novel synthetic task in which a Transformer is trained from scratch to internalize a "structured" knowledge graph. The structure enforces relationships between entities of the graph, such that editing a factual association has "trickling effects" on other entities (e.g., altering X's parent is Y to Z affects who X's siblings' parent is). Through evaluations of edited models on this task, we show that KE inadvertently affects representations of entities beyond the targeted one, distorting relevant structures that allow a model to infer unseen knowledge about an entity. We call this phenomenon representation shattering and demonstrate that it degrades models' factual recall and reasoning performance. We further corroborate our findings in naturalistic settings with pre-trained Llama and Mamba models as well. Overall, our work yields a precise mechanistic hypothesis to explain why KE has adverse effects on model abilities.
Authors: Adit Jain, Soumyabrata Pal, Sunav Choudhary, Ramasuri Narayanam, Harshita Chopra, Vikram Krishnamurthy
Abstract: We investigate the high-dimensional sparse linear bandits problem in a data-poor regime where the time horizon is much smaller than the ambient dimension and number of arms. We study the setting under the additional blocking constraint where each unique arm can be pulled only once. The blocking constraint is motivated by practical applications in personalized content recommendation and identification of data points to improve annotation efficiency for complex learning tasks. With mild assumptions on the arms, our proposed online algorithm (BSLB) achieves a regret guarantee of $\widetilde{\mathsf{O}}((1+\beta_k)^2k^{\frac{2}{3}} \mathsf{T}^{\frac{2}{3}})$ where the parameter vector has an (unknown) relative tail $\beta_k$ -- the ratio of $\ell_1$ norm of the top-$k$ and remaining entries of the parameter vector. To this end, we show novel offline statistical guarantees of the lasso estimator for the linear model that is robust to the sparsity modeling assumption. Finally, we propose a meta-algorithm (C-BSLB) based on corralling that does not need knowledge of optimal sparsity parameter $k$ at minimal cost to regret. Our experiments on multiple real-world datasets demonstrate the validity of our algorithms and theoretical framework.
Authors: Yihua Shao, Yan Gu, Siyu Chen, Haiyang Liu, Zixian Zhu, Zijian Ling, Minxi Yan, Ziyang Yan, Chenyu Zhang, Michele Magno, Haotong Qin, Yan Wang, Jingcai Guo, Ling Shao, Hao Tang
Abstract: Large language models (LLMs) show impressive performance in solving complex language tasks. However, its large number of parameters presents significant challenges for the deployment. So, compressing LLMs to low bits can enable to deploy on resource-constrained devices. To address this problem, we propose gradient-aware weight quantization (GWQ), the first quantization approach for low-bit weight quantization that leverages gradients to localize outliers, requiring only a minimal amount of calibration data for outlier detection. GWQ retains the top 1\% outliers preferentially at FP16 precision, while the remaining non-outlier weights are stored in a low-bit. We widely evaluate GWQ on different task include language modeling, grounding detection, massive multitask language understanding and vision-language question and answering. Results show that models quantified by GWQ performs better than other quantization method. During quantization process, GWQ only need one calibration set to realize effective quant. Also, GWQ achieves 1.2x inference speedup in comparison to the original model and effectively reduces the inference memory.
Authors: Guangyi Wang, Wei Peng, Lijiang Li, Wenyu Chen, Yuren Cai, Songzhi Su
Abstract: While powerful for generation, Diffusion Probabilistic Models (DPMs) face slow sampling challenges, for which various distillation-based methods have been proposed. However, they typically require significant additional training costs and model parameter storage, limiting their practicality. In this work, we propose PCA-based Adaptive Search (PAS), which optimizes existing solvers for DPMs with minimal additional costs. Specifically, we first employ PCA to obtain a few basis vectors to span the high-dimensional sampling space, which enables us to learn just a set of coordinates to correct the sampling direction; furthermore, based on the observation that the cumulative truncation error exhibits an ``S"-shape, we design an adaptive search strategy that further enhances the sampling efficiency and reduces the number of stored parameters to approximately 10. Extensive experiments demonstrate that PAS can significantly enhance existing fast solvers in a plug-and-play manner with negligible costs. E.g., on CIFAR10, PAS optimizes DDIM's FID from 15.69 to 4.37 (NFE=10) using only 12 parameters and sub-minute training on a single A100 GPU. Code is available at https://github.com/onefly123/PAS.
Authors: Guoguo Ai, Guansong Pang, Hezhe Qiao, Yuan Gao, Hui Yan
Abstract: Graph Transformers (GTs) have demonstrated remarkable performance in graph representation learning over popular graph neural networks (GNNs). However, self--attention, the core module of GTs, preserves only low-frequency signals in graph features, leading to ineffectiveness in capturing other important signals like high-frequency ones. Some recent GT models help alleviate this issue, but their flexibility and expressiveness are still limited since the filters they learn are fixed on predefined graph spectrum or spectral order. To tackle this challenge, we propose a Graph Fourier Kolmogorov-Arnold Transformer (GrokFormer), a novel GT model that learns highly expressive spectral filters with adaptive graph spectrum and spectral order through a Fourier series modeling over learnable activation functions. We demonstrate theoretically and empirically that the proposed GrokFormer filter offers better expressiveness than other spectral methods. Comprehensive experiments on 10 real-world node classification datasets across various domains, scales, and graph properties, as well as 5 graph classification datasets, show that GrokFormer outperforms state-of-the-art GTs and GNNs. Our code is available at https://github.com/GGA23/GrokFormer
Authors: Aladin Djuhera, Amin Seffo, Vlad C. Andrei, Holger Boche, Walid Saad
Abstract: Path planning under wireless performance constraints is a complex challenge in robot navigation. However, naively incorporating such constraints into classical planning algorithms often incurs prohibitive search costs. In this paper, we propose SCoTT, a wireless-aware path planning framework that leverages vision-language models (VLMs) to co-optimize average path gains and trajectory length using wireless heatmap images and ray-tracing data from a digital twin (DT). At the core of our framework is Strategic Chain-of-Thought Tasking (SCoTT), a novel prompting paradigm that decomposes the exhaustive search problem into structured subtasks, each solved via chain-of-thought prompting. To establish strong baselines, we compare classical A* and wireless-aware extensions of it, and derive DP-WA*, an optimal, iterative dynamic programming algorithm that incorporates all path gains and distance metrics from the DT, but at significant computational cost. In extensive experiments, we show that SCoTT achieves path gains within 2% of DP-WA* while consistently generating shorter trajectories. Moreover, SCoTT's intermediate outputs can be used to accelerate DP-WA* by reducing its search space, saving up to 62% in execution time. We validate our framework using four VLMs, demonstrating effectiveness across both large and small models, thus making it applicable to a wide range of compact models at low inference cost. We also show the practical viability of our approach by deploying SCoTT as a ROS node within Gazebo simulations. Finally, we discuss data acquisition pipelines, compute requirements, and deployment considerations for VLMs in 6G-enabled DTs, underscoring the potential of natural language interfaces for wireless-aware navigation in real-world applications.
Authors: Jingzhe Liu, Haitao Mao, Zhikai Chen, Bingheng Li, Wenqi Fan, Mingxuan Ju, Tong Zhao, Neil Shah, Jiliang Tang
Abstract: Graph Neural Networks (GNNs) have emerged as a powerful tool to capture intricate network patterns, achieving success across different domains. However, existing GNNs require careful domain-specific architecture designs and training from scratch on each dataset, leading to an expertise-intensive process with difficulty in generalizing across graphs from different domains. Therefore, it can be hard for practitioners to infer which GNN model can generalize well to graphs from their domains. To address this challenge, we propose a novel cross-domain pretraining framework, "one model for one graph," which overcomes the limitations of previous approaches that failed to use a single GNN to capture diverse graph patterns across domains with significant gaps. Specifically, we pretrain a bank of expert models, with each one corresponding to a specific dataset. When inferring to a new graph, gating functions choose a subset of experts to effectively integrate prior model knowledge while avoiding negative transfer. Extensive experiments consistently demonstrate the superiority of our proposed method on both link prediction and node classification tasks.
Authors: Seockbean Song, Youngsik Yoon, Siwei Wang, Wei Chen, Jungseul Ok
Abstract: Combinatorial online learning is a fundamental task for selecting the optimal action (or super arm) as a combination of base arms in sequential interactions with systems providing stochastic rewards. It is applicable to diverse domains such as robotics, social advertising, network routing, and recommendation systems. In many real-world scenarios, we often encounter rising rewards, where playing a base arm not only provides an instantaneous reward but also contributes to the enhancement of future rewards, e.g., robots enhancing proficiency through practice and social influence strengthening in the history of successful recommendations. Moreover, the enhancement of a single base arm may affect multiple super arms that include it, introducing complex dependencies that are not captured by existing rising bandit models. To address this, we introduce the Combinatorial Rising Bandit (CRB) framework and propose a provably efficient algorithm, Combinatorial Rising Upper Confidence Bound (CRUCB). We establish an upper bound on regret CRUCB and show that it is nearly tight by deriving a matching lower bound. In addition, we empirically demonstrate the effectiveness of CRUCB not only in synthetic environments but also in realistic applications of deep reinforcement learning.
Authors: Yuyang Wang, Anurag Ranjan, Josh Susskind, Miguel Angel Bautista
Abstract: Flow matching models have emerged as a powerful method for generative modeling on domains like images or videos, and even on irregular or unstructured data like 3D point clouds or even protein structures. These models are commonly trained in two stages: first, a data compressor is trained, and in a subsequent training stage a flow matching generative model is trained in the latent space of the data compressor. This two-stage paradigm sets obstacles for unifying models across data domains, as hand-crafted compressors architectures are used for different data modalities. To this end, we introduce INRFlow, a domain-agnostic approach to learn flow matching transformers directly in ambient space. Drawing inspiration from INRs, we introduce a conditionally independent point-wise training objective that enables INRFlow to make predictions continuously in coordinate space. Our empirical results demonstrate that INRFlow effectively handles different data modalities such as images, 3D point clouds and protein structure data, achieving strong performance in different domains and outperforming comparable approaches. INRFlow is a promising step towards domain-agnostic flow matching generative models that can be trivially adopted in different data domains.
Authors: Ziang Zhou, Zhihao Ding, Jieming Shi, Qing Li, Shiqi Shen
Abstract: Graph Neural Networks (GNNs) are pivotal in graph-based learning, particularly excelling in node classification. However, their scalability is hindered by the need for multi-hop data during inference, limiting their application in latency-sensitive scenarios. Recent efforts to distill GNNs into multi-layer perceptrons (MLPs) for faster inference often underutilize the layer-level insights of GNNs. In this paper, we present TINED, a novel approach that distills GNNs to MLPs on a layer-by-layer basis using Teacher Injection and Dirichlet Energy Distillation techniques. We focus on two key operations in GNN layers: feature transformation (FT) and graph propagation (GP). We recognize that FT is computationally equivalent to a fully-connected (FC) layer in MLPs. Thus, we propose directly transferring teacher parameters from an FT in a GNN to an FC layer in the student MLP, enhanced by fine-tuning. In TINED, the FC layers in an MLP replicate the sequence of FTs and GPs in the GNN. We also establish a theoretical bound for GP approximation. Furthermore, we note that FT and GP operations in GNN layers often exhibit opposing smoothing effects: GP is aggressive, while FT is conservative. Using Dirichlet energy, we develop a DE ratio to measure these effects and propose Dirichlet Energy Distillation to convey these characteristics from GNN layers to MLP layers. Extensive experiments show that TINED outperforms GNNs and leading distillation methods across various settings and seven datasets. Source code are available at https://github.com/scottjiao/TINED_ICML25/.
Authors: Jaeheun Jung, Jaehyuk Lee, Changhae Jung, Hanyoung Kim, Bosung Jung, Donghun Lee
Abstract: Shock waves caused by earthquakes can be devastating. Generating realistic earthquake-caused ground motion waveforms help reducing losses in lives and properties, yet generative models for the task tend to generate subpar waveforms. We present High-fidelity Earthquake Groundmotion Generation System (HEGGS) and demonstrate its superior performance using earthquakes from North American, East Asian, and European regions. HEGGS exploits the intrinsic characteristics of earthquake dataset and learns the waveforms using an end-to-end differentiable generator containing conditional latent diffusion model and hi-fidelity waveform construction model. We show the learning efficiency of HEGGS by training it on a single GPU machine and validate its performance using earthquake databases from North America, East Asia, and Europe, using diverse criteria from waveform generation tasks and seismology. Once trained, HEGGS can generate three dimensional E-N-Z seismic waveforms with accurate P/S phase arrivals, envelope correlation, signal-to-noise ratio, GMPE analysis, frequency content analysis, and section plot analysis.
Authors: Zhuohuan Hu, Richard Yu, Zizhou Zhang, Haoran Zheng, Qianying Liu, Yining Zhou
Abstract: This paper leverages machine learning algorithms to forecast and analyze financial time series. The process begins with a denoising autoencoder to filter out random noise fluctuations from the main contract price data. Then, one-dimensional convolution reduces the dimensionality of the filtered data and extracts key information. The filtered and dimensionality-reduced price data is fed into a GANs network, and its output serve as input of a fully connected network. Through cross-validation, a model is trained to capture features that precede large price fluctuations. The model predicts the likelihood and direction of significant price changes in real-time price sequences, placing trades at moments of high prediction accuracy. Empirical results demonstrate that using autoencoders and convolution to filter and denoise financial data, combined with GANs, achieves a certain level of predictive performance, validating the capabilities of machine learning algorithms to discover underlying patterns in financial sequences. Keywords - CNN;GANs; Cryptocurrency; Prediction.
Authors: Chaoqi Wang, Zhuokai Zhao, Yibo Jiang, Zhaorun Chen, Chen Zhu, Yuxin Chen, Jiayi Liu, Lizhu Zhang, Xiangjun Fan, Hao Ma, Sinong Wang
Abstract: Recent advances in large language models (LLMs) have demonstrated significant progress in performing complex tasks. While Reinforcement Learning from Human Feedback (RLHF) has been effective in aligning LLMs with human preferences, it is susceptible to spurious correlations in reward modeling. Consequently, it often introduces biases-such as length bias, sycophancy, conceptual bias, and discrimination-that hinder the model's ability to capture true causal relationships. To address this, we propose a novel causal reward modeling approach that integrates causality to mitigate these spurious correlations. Our method enforces counterfactual invariance, ensuring reward predictions remain consistent when irrelevant variables are altered. Through experiments on both synthetic and real-world datasets, we show that our approach mitigates various types of spurious correlations effectively, resulting in more reliable and fair alignment of LLMs with human preferences. As a drop-in enhancement to the existing RLHF workflow, our causal reward modeling provides a practical way to improve the trustworthiness and fairness of LLM finetuning.
Authors: Zizhou Zhang, Xinshi Li, Yu Cheng, Zhenrui Chen, Qianying Liu
Abstract: Credit risk management within supply chains has emerged as a critical research area due to its significant implications for operational stability and financial sustainability. The intricate interdependencies among supply chain participants mean that credit risks can propagate across networks, with impacts varying by industry. This study explores the application of Generative Adversarial Networks (GANs) to enhance credit risk identification in supply chains. GANs enable the generation of synthetic credit risk scenarios, addressing challenges related to data scarcity and imbalanced datasets. By leveraging GAN-generated data, the model improves predictive accuracy while effectively capturing dynamic and temporal dependencies in supply chain data. The research focuses on three representative industries-manufacturing (steel), distribution (pharmaceuticals), and services (e-commerce) to assess industry-specific credit risk contagion. Experimental results demonstrate that the GAN-based model outperforms traditional methods, including logistic regression, decision trees, and neural networks, achieving superior accuracy, recall, and F1 scores. The findings underscore the potential of GANs in proactive risk management, offering robust tools for mitigating financial disruptions in supply chains. Future research could expand the model by incorporating external market factors and supplier relationships to further enhance predictive capabilities. Keywords- Generative Adversarial Networks (GANs); Supply Chain Risk; Credit Risk Identification; Machine Learning; Data Augmentation
Authors: Ayush Mohanty, Nazal Mohamed, Paritosh Ramanan, Nagi Gebraeel
Abstract: Advanced sensors and IoT devices have improved the monitoring and control of complex industrial enterprises. They have also created an interdependent fabric of geographically distributed process operations (clients) across these enterprises. Granger causality is an effective approach to detect and quantify interdependencies by examining how one client's state affects others over time. Understanding these interdependencies captures how localized events, such as faults and disruptions, can propagate throughout the system, possibly causing widespread operational impacts. However, the large volume and complexity of industrial data pose challenges in modeling these interdependencies. This paper develops a federated approach to learning Granger causality. We utilize a linear state space system framework that leverages low-dimensional state estimates to analyze interdependencies. This addresses bandwidth limitations and the computational burden commonly associated with centralized data processing. We propose augmenting the client models with the Granger causality information learned by the server through a Machine Learning (ML) function. We examine the co-dependence between the augmented client and server models and reformulate the framework as a standalone ML algorithm providing conditions for its sublinear and linear convergence rates. We also study the convergence of the framework to a centralized oracle model. Moreover, we include a differential privacy analysis to ensure data security while preserving causal insights. Using synthetic data, we conduct comprehensive experiments to demonstrate the robustness of our approach to perturbations in causality, the scalability to the size of communication, number of clients, and the dimensions of raw data. We also evaluate the performance on two real-world industrial control system datasets by reporting the volume of data saved by decentralization.
Authors: Shichang Zhang, Tessa Han, Usha Bhalla, Himabindu Lakkaraju
Abstract: The increasing complexity of AI systems has made understanding their behavior critical. Numerous interpretability methods have been developed to attribute model behavior to three key aspects: input features, training data, and internal model components, which emerged from explainable AI, data-centric AI, and mechanistic interpretability, respectively. However, these attribution methods are studied and applied rather independently, resulting in a fragmented landscape of methods and terminology. This position paper argues that feature, data, and component attribution methods share fundamental similarities, and a unified view of them benefits both interpretability and broader AI research. To this end, we first analyze popular methods for these three types of attributions and present a unified view demonstrating that these seemingly distinct methods employ similar techniques (such as perturbations, gradients, and linear approximations) over different aspects and thus differ primarily in their perspectives rather than techniques. Then, we demonstrate how this unified view enhances understanding of existing attribution methods, highlights shared concepts and evaluation criteria among these methods, and leads to new research directions both in interpretability research, by addressing common challenges and facilitating cross-attribution innovation, and in AI more broadly, with applications in model editing, steering, and regulation.
Authors: Ruigang Wang, Krishnamurthy Dvijotham, Ian R. Manchester
Abstract: In this work, we propose norm-bounded low-rank adaptation (NB-LoRA) for parameter-efficient fine tuning. NB-LoRA is a novel parameterization of low-rank weight adaptations that admits explicit bounds on each singular value of the adaptation matrix, which can thereby satisfy any prescribed unitarily invariant norm bound, including the Schatten norms (e.g., nuclear, Frobenius, spectral norm). The proposed parameterization is unconstrained, smooth, and complete, i.e. it covers all matrices satisfying the prescribed rank and singular-value bounds. Comparative experiments on large language models show that NB-LoRA achieves superior adaptation performance and faster training over a range of models, tasks and ranks. Vision fine-tuning experiments show that NB-LoRA can achieve strong adaptation performance while avoiding model catastrophic forgetting, and compared to existing approaches it is substantially more robust to a hyper-parameters such as including adaptation rank, learning rate and number of training epochs.
Authors: Xutong Liu, Xiangxiang Dai, Jinhang Zuo, Siwei Wang, Carlee Joe-Wong, John C. S. Lui, Wei Chen
Abstract: The combinatorial multi-armed bandit (CMAB) is a fundamental sequential decision-making framework, extensively studied over the past decade. However, existing work primarily focuses on the online setting, overlooking the substantial costs of online interactions and the readily available offline datasets. To overcome these limitations, we introduce Off-CMAB, the first offline learning framework for CMAB. Central to our framework is the combinatorial lower confidence bound (CLCB) algorithm, which combines pessimistic reward estimations with combinatorial solvers. To characterize the quality of offline datasets, we propose two novel data coverage conditions and prove that, under these conditions, CLCB achieves a near-optimal suboptimality gap, matching the theoretical lower bound up to a logarithmic factor. We validate Off-CMAB through practical applications, including learning to rank, large language model (LLM) caching, and social influence maximization, showing its ability to handle nonlinear reward functions, general feedback models, and out-of-distribution action samples that excludes optimal or even feasible actions. Extensive experiments on synthetic and real-world datasets further highlight the superior performance of CLCB.
Authors: Binchi Zhang, Zaiyi Zheng, Zhengzhang Chen, Jundong Li
Abstract: Symmetry in the parameter space of deep neural networks (DNNs) has proven beneficial for various deep learning applications. A well-known example is the permutation symmetry in Multi-Layer Perceptrons (MLPs), where permuting the rows of weight matrices in one layer and applying the inverse permutation to adjacent layers yields a functionally equivalent model. While permutation symmetry fully characterizes the equivalence set for MLPs, its discrete nature limits its utility for transformers. In this paper, we introduce rotation symmetry, a novel form of parameter space symmetry for transformers that generalizes permutation symmetry by rotating parameter matrices in self-attention layers. Unlike permutation symmetry, rotation symmetry operates in a continuous domain, thereby significantly expanding the equivalence set for transformers. Based on this property, we propose a theoretically optimal parameter matching algorithm as a plug-and-play module to enhance model fusion. We evaluate our approach using pre-trained transformers across diverse natural language and vision tasks. Experimental results demonstrate that our rotation symmetry-based matching algorithm substantially improves model fusion, highlighting the potential of parameter space symmetry to facilitate model fusion. Our code is available on https://github.com/zhengzaiyi/RotationSymmetry.
Authors: Jijia Liu, Feng Gao, Qingmin Liao, Chao Yu, Yu Wang
Abstract: Reinforcement learning (RL) for continuous control often requires large amounts of online interaction data. Value-based RL methods can mitigate this burden by offering relatively high sample efficiency. Some studies further enhance sample efficiency by incorporating offline demonstration data to "kick-start" training, achieving promising results in continuous control. However, they typically compute the Q-function independently for each action dimension, neglecting interdependencies and making it harder to identify optimal actions when learning from suboptimal data, such as non-expert demonstration and online-collected data during the training process. To address these issues, we propose Auto-Regressive Soft Q-learning (ARSQ), a value-based RL algorithm that models Q-values in a coarse-to-fine, auto-regressive manner. First, ARSQ decomposes the continuous action space into discrete spaces in a coarse-to-fine hierarchy, enhancing sample efficiency for fine-grained continuous control tasks. Next, it auto-regressively predicts dimensional action advantages within each decision step, enabling more effective decision-making in continuous control tasks. We evaluate ARSQ on two continuous control benchmarks, RLBench and D4RL, integrating demonstration data into online training. On D4RL, which includes non-expert demonstrations, ARSQ achieves an average $1.62\times$ performance improvement over SOTA value-based baseline. On RLBench, which incorporates expert demonstrations, ARSQ surpasses various baselines, demonstrating its effectiveness in learning from suboptimal online-collected data. Project page is at https://sites.google.com/view/ar-soft-q
Authors: Chuqi Chen, Yahong Yang, Yang Xiang, Wenrui Hao
Abstract: Solving partial differential equations (PDEs) using neural networks has become a central focus in scientific machine learning. Training neural networks for singularly perturbed problems is particularly challenging due to certain parameters in the PDEs that introduce near-singularities in the loss function. In this study, we overcome this challenge by introducing a novel method based on homotopy dynamics to effectively manipulate these parameters. From a theoretical perspective, we analyze the effects of these parameters on training difficulty in these singularly perturbed problems and establish the convergence of the proposed homotopy dynamics method. Experimentally, we demonstrate that our approach significantly accelerates convergence and improves the accuracy of these singularly perturbed problems. These findings present an efficient optimization strategy leveraging homotopy dynamics, offering a robust framework to extend the applicability of neural networks for solving singularly perturbed differential equations.
Authors: Roman Tarasov, Petr Mokrov, Milena Gazdieva, Evgeny Burnaev, Alexander Korotin
Abstract: Neural network-based optimal transport (OT) is a recent and fruitful direction in the generative modeling community. It finds its applications in various fields such as domain translation, image super-resolution, computational biology and others. Among the existing OT approaches, of considerable interest are adversarial minimax solvers based on semi-dual formulations of OT problems. While promising, these methods lack theoretical investigation from a statistical learning perspective. Our work fills this gap by establishing upper bounds on the generalization error of an approximate OT map recovered by the minimax quadratic OT solver. Importantly, the bounds we derive depend solely on some standard statistical and mathematical properties of the considered functional classes (neural nets). While our analysis focuses on the quadratic OT, we believe that similar bounds could be derived for general OT case, paving the promising direction for future research.
Authors: Tianlin Zhang, En Yu, Yi Shao, Jiande Sun
Abstract: Multimodal fake news detection has garnered significant attention due to its profound implications for social security. While existing approaches have contributed to understanding cross-modal consistency, they often fail to leverage modal-specific representations and explicit discrepant features. To address these limitations, we propose a Multimodal Inverse Attention Network (MIAN), a novel framework that explores intrinsic discriminative features based on news content to advance fake news detection. Specifically, MIAN introduces a hierarchical learning module that captures diverse intra-modal relationships through local-to-global and local-to-local interactions, thereby generating enhanced unimodal representations to improve the identification of fake news at the intra-modal level. Additionally, a cross-modal interaction module employs a co-attention mechanism to establish and model dependencies between the refined unimodal representations, facilitating seamless semantic integration across modalities. To explicitly extract inconsistency features, we propose an inverse attention mechanism that effectively highlights the conflicting patterns and semantic deviations introduced by fake news in both intra- and inter-modality. Extensive experiments on benchmark datasets demonstrate that MIAN significantly outperforms state-of-the-art methods, underscoring its pivotal contribution to advancing social security through enhanced multimodal fake news detection.
Authors: Jan Schuchardt, Mina Dalirrooyfard, Jed Guzelkabaagac, Anderson Schneider, Yuriy Nevmyvaka, Stephan G\"unnemann
Abstract: Many forms of sensitive data, such as web traffic, mobility data, or hospital occupancy, are inherently sequential. The standard method for training machine learning models while ensuring privacy for units of sensitive information, such as individual hospital visits, is differentially private stochastic gradient descent (DP-SGD). However, we observe in this work that the formal guarantees of DP-SGD are incompatible with time-series-specific tasks like forecasting, since they rely on the privacy amplification attained by training on small, unstructured batches sampled from an unstructured dataset. In contrast, batches for forecasting are generated by (1) sampling sequentially structured time series from a dataset, (2) sampling contiguous subsequences from these series, and (3) partitioning them into context and ground-truth forecast windows. We theoretically analyze the privacy amplification attained by this structured subsampling to enable the training of forecasting models with sound and tight event- and user-level privacy guarantees. Towards more private models, we additionally prove how data augmentation amplifies privacy in self-supervised training of sequence models. Our empirical evaluation demonstrates that amplification by structured subsampling enables the training of forecasting models with strong formal privacy guarantees.
Authors: Alessio Russo, Aldo Pacchiano
Abstract: We study the policy evaluation problem in an online multi-reward multi-policy discounted setting, where multiple reward functions must be evaluated simultaneously for different policies. We adopt an $(\epsilon,\delta)$-PAC perspective to achieve $\epsilon$-accurate estimates with high confidence across finite or convex sets of rewards, a setting that has not been investigated in the literature. Building on prior work on Multi-Reward Best Policy Identification, we adapt the MR-NaS exploration scheme to jointly minimize sample complexity for evaluating different policies across different reward sets. Our approach leverages an instance-specific lower bound revealing how the sample complexity scales with a measure of value deviation, guiding the design of an efficient exploration policy. Although computing this bound entails a hard non-convex optimization, we propose an efficient convex approximation that holds for both finite and convex reward sets. Experiments in tabular domains demonstrate the effectiveness of this adaptive exploration scheme.
Authors: Adibvafa Fallahpour, Jun Ma, Alif Munim, Hongwei Lyu, Bo Wang
Abstract: Chest X-rays (CXRs) play an integral role in driving critical decisions in disease management and patient care. While recent innovations have led to specialized models for various CXR interpretation tasks, these solutions often operate in isolation, limiting their practical utility in clinical practice. We present MedRAX, the first versatile AI agent that seamlessly integrates state-of-the-art CXR analysis tools and multimodal large language models into a unified framework. MedRAX dynamically leverages these models to address complex medical queries without requiring additional training. To rigorously evaluate its capabilities, we introduce ChestAgentBench, a comprehensive benchmark containing 2,500 complex medical queries across 7 diverse categories. Our experiments demonstrate that MedRAX achieves state-of-the-art performance compared to both open-source and proprietary models, representing a significant step toward the practical deployment of automated CXR interpretation systems. Data and code have been publicly available at https://github.com/bowang-lab/MedRAX
Authors: Chen-Xiao Gao, Chenyang Wu, Mingjun Cao, Chenjun Xiao, Yang Yu, Zongzhang Zhang
Abstract: Behavior regularization, which constrains the policy to stay close to some behavior policy, is widely used in offline reinforcement learning (RL) to manage the risk of hazardous exploitation of unseen actions. Nevertheless, existing literature on behavior-regularized RL primarily focuses on explicit policy parameterizations, such as Gaussian policies. Consequently, it remains unclear how to extend this framework to more advanced policy parameterizations, such as diffusion models. In this paper, we introduce BDPO, a principled behavior-regularized RL framework tailored for diffusion-based policies, thereby combining the expressive power of diffusion policies and the robustness provided by regularization. The key ingredient of our method is to calculate the Kullback-Leibler (KL) regularization analytically as the accumulated discrepancies in reverse-time transition kernels along the diffusion trajectory. By integrating the regularization, we develop an efficient two-time-scale actor-critic RL algorithm that produces the optimal policy while respecting the behavior constraint. Comprehensive evaluations conducted on synthetic 2D tasks and continuous control tasks from the D4RL benchmark validate its effectiveness and superior performance.
Authors: Xiao Li, Zekai Zhang, Xiang Li, Siyi Chen, Zhihui Zhu, Peng Wang, Qing Qu
Abstract: Diffusion models, though originally designed for generative tasks, have demonstrated impressive self-supervised representation learning capabilities. A particularly intriguing phenomenon in these models is the emergence of unimodal representation dynamics, where the quality of learned features peaks at an intermediate noise level. In this work, we conduct a comprehensive theoretical and empirical investigation of this phenomenon. Leveraging the inherent low-dimensionality structure of image data, we theoretically demonstrate that the unimodal dynamic emerges when the diffusion model successfully captures the underlying data distribution. The unimodality arises from an interplay between denoising strength and class confidence across noise scales. Empirically, we further show that, in classification tasks, the presence of unimodal dynamics reliably indicates generalization: it emerges when the model generalizes and gradually transitions to a monotonically decreasing curve as the model begins to memorize the training data.
Authors: Annie Marsden, Elad Hazan
Abstract: We study the problem of preconditioning in the setting of sequential prediction. From the theoretical lens of linear dynamical systems, we show that applying a convolution to the input sequence translates to applying a polynomial to the unknown transition matrix in the hidden space. With this insight, we develop a novel preconditioning method that convolves the input sequence with the coefficients of the Chebyshev or Legendre polynomials. We formally prove that this improves the regret of two distinct prediction methods. Moreover, using this preconditioning technique on either method gives the first sublinear regret bounds that are also hidden dimension independent (up to logarithmic factors) even when the hidden transition matrix is asymmetric. From rigorous experiments on synthetic data we show that our simple preconditioning method generalizes to both 1) settings where the data is not from a linear dynamical system and 2) a broad range of learning algorithms, including recurrent neural networks.
Authors: Jasraj Singh, Keyue Jiang, Brooks Paige, Laura Toni
Abstract: Message Passing Neural Networks (MPNNs) are a class of Graph Neural Networks (GNNs) that propagate information across the graph via local neighborhoods. The scheme gives rise to two key challenges: over-smoothing and over-squashing. While several Dropout-style algorithms, such as DropEdge and DropMessage, have successfully addressed over-smoothing, their impact on over-squashing remains largely unexplored. This represents a critical gap in the literature, as failure to mitigate over-squashing would make these methods unsuitable for long-range tasks -- the intended use case of deep MPNNs. In this work, we study the aforementioned algorithms, and closely related edge-dropping algorithms -- DropNode, DropAgg and DropGNN -- in the context of over-squashing. We present theoretical results showing that DropEdge-variants reduce sensitivity between distant nodes, limiting their suitability for long-range tasks. To address this, we introduce DropSens, a sensitivity-aware variant of DropEdge that explicitly controls the proportion of information lost due to edge-dropping, thereby increasing sensitivity to distant nodes despite dropping the same number of edges. Our experiments on long-range synthetic and real-world datasets confirm the predicted limitations of existing edge-dropping and feature-dropping methods. Moreover, DropSens consistently outperforms graph rewiring techniques designed to mitigate over-squashing, suggesting that simple, targeted modifications can substantially improve a model's ability to capture long-range interactions. Our conclusions highlight the need to re-evaluate and re-design existing methods for training deep GNNs, with a renewed focus on modelling long-range interactions.
Authors: Yanran Wu, Inez Hua, Yi Ding
Abstract: Large language models (LLMs) offer powerful capabilities but come with significant environmental impact, particularly in carbon emissions. Existing studies benchmark carbon emissions but lack a standardized basis for comparison across different model configurations. To address this, we introduce the concept of functional unit (FU) as a standardized basis and develop FUEL, the first FU-based framework for evaluating LLM serving's environmental impact. Through three case studies, we uncover key insights and trade-offs in reducing carbon emissions by optimizing model size, quantization strategy, and hardware choice, paving the way for more sustainable LLM serving. The code is available at https://github.com/jojacola/FUEL.
Authors: Jiaqi Wang, Xinyi Wu, James Cheng, Yifei Wang
Abstract: Deep graph neural networks (GNNs) often suffer from oversmoothing, where node representations become overly homogeneous with increasing depth. While techniques like normalization, residual connections, and edge dropout have been proposed to mitigate oversmoothing, they are typically developed independently, with limited theoretical understanding of their underlying mechanisms. In this work, we present a unified theoretical perspective based on the framework of signed graphs, showing that many existing strategies implicitly introduce negative edges that alter message-passing to resist oversmoothing. However, we show that merely adding negative edges in an unstructured manner is insufficient-the asymptotic behavior of signed propagation depends critically on the strength and organization of positive and negative edges. To address this limitation, we leverage the theory of structural balance, which promotes stable, cluster-preserving dynamics by connecting similar nodes with positive edges and dissimilar ones with negative edges. We propose Structural Balanced Propagation (SBP), a plug-and-play method that assigns signed edges based on either labels or feature similarity to explicitly enhance structural balance in the constructed signed graphs. Experiments on nine benchmarks across both homophilic and heterophilic settings demonstrate that SBP consistently improves classification accuracy and mitigates oversmoothing, even at depths of up to 300 layers. Our results provide a principled explanation for prior oversmoothing remedies and introduce a new direction for signed message-passing design in deep GNNs.
Authors: Jichan Chung, Irene Y. Chen
Abstract: The high cost of data labeling presents a major barrier to deploying machine learning systems at scale. Semi-supervised learning (SSL) mitigates this challenge by utilizing unlabeled data alongside limited labeled examples, while the emergence of foundation models (FMs) offers powerful zero-shot capabilities that can further reduce labeling cost. However, directly fine-tuning large FMs is often impractical in resource-constrained settings, and na\"ively using their pseudo-labels for unlabeled data can degrade performance due to its unreliablity or domain mismatch with target task. In this work, we introduce ZeroMatch, a novel SSL framework that integrates knowledge distillation with consistency-based learning to jointly leverage labeled data, unlabeled data, and pseudo-labels from FMs. ZeroMatch enables training compact student models using only FM inference, making it suitable for low-resource environments such as personal devices with limited compute. Experiments on six vision and language classification benchmarks show that ZeroMatch consistently outperforms standard SSL and zero-shot augmented methods, demonstrating its effectiveness and robustness across a range of foundation model qualities.
Authors: Sifan Zhou, Shuo Wang, Zhihang Yuan, Mingjia Shi, Yuzhang Shang, Dawei Yang
Abstract: Large Language Models (LLMs) fine-tuning technologies have achieved remarkable results. However, traditional LLM fine-tuning approaches face significant challenges: they require large Floating Point (FP) computation, raising privacy concerns when handling sensitive data, and are impractical for resource-constrained edge devices. While Parameter-Efficient Fine-Tuning (PEFT) techniques reduce trainable parameters, their reliance on floating-point arithmetic creates fundamental incompatibilities with edge hardware. In this work, we introduce a novel framework for on-device LLM fine-tuning that eliminates the need for floating-point operations in both inference and training, named GSQ-Tuning. At its core is the Group-Shared Exponents Integer format, which efficiently represents model parameters in integer format using shared exponents among parameter groups. When combined with LoRA-like adapters, this enables fully integer-based fine-tuning that is both memory and compute efficient. We demonstrate that our approach achieves accuracy comparable to BF16-based fine-tuning while significantly reducing 1.85x memory usage. Moreover, compared to FP8, our method can reduce 5x power consumption and 11x chip area with same performance, making large-scale model adaptation feasible on edge devices.
Authors: Tassallah Abdullahi, Ioanna Gemou, Nihal V. Nayak, Ghulam Murtaza, Stephen H. Bach, Carsten Eickhoff, Ritambhara Singh
Abstract: Biomedical knowledge graphs (KGs) encode rich, structured information critical for drug discovery tasks, but extracting meaningful insights from large-scale KGs remains challenging due to their complex structure. Existing biomedical subgraph retrieval methods are tailored for graph neural networks (GNNs), limiting compatibility with other paradigms, including large language models (LLMs). We introduce K-Paths, a model-agnostic retrieval framework that extracts structured, diverse, and biologically meaningful multi-hop paths from dense biomedical KGs. These paths enable the prediction of unobserved drug-drug and drug-disease interactions, including those involving entities not seen during training, thus supporting inductive reasoning. K-Paths is training-free and employs a diversity-aware adaptation of Yen's algorithm to extract the K shortest loopless paths between entities in a query, prioritizing biologically relevant and relationally diverse connections. These paths serve as concise, interpretable reasoning chains that can be directly integrated with LLMs or GNNs to improve generalization, accuracy, and enable explainable inference. Experiments on benchmark datasets show that K-Paths improves zero-shot reasoning across state-of-the-art LLMs. For instance, Tx-Gemma 27B improves by 19.8 and 4.0 F1 points on interaction severity prediction and drug repurposing tasks, respectively. Llama 70B achieves gains of 8.5 and 6.2 points on the same tasks. K-Paths also boosts the training efficiency of EmerGNN, a state-of-the-art GNN, by reducing the KG size by 90% while maintaining predictive performance. Beyond efficiency, K-Paths bridges the gap between KGs and LLMs, enabling scalable and explainable LLM-augmented scientific discovery. We release our code and the retrieved paths as a benchmark for inductive reasoning.
Authors: Hanlin Wang, Jian Wang, Chak Tou Leong, Wenjie Li
Abstract: Large language model (LLM)-based agents have shown promise in tackling complex tasks by interacting dynamically with the environment. Existing work primarily focuses on behavior cloning from expert demonstrations or preference learning through exploratory trajectory sampling. However, these methods often struggle to address long-horizon tasks, where suboptimal actions accumulate step by step, causing agents to deviate from correct task trajectories. To address this, we highlight the importance of timely calibration and the need to automatically construct calibration trajectories for training agents. We propose Step-Level Trajectory Calibration (STeCa), a novel framework for LLM agent learning. Specifically, STeCa identifies suboptimal actions through a step-level reward comparison during exploration. It constructs calibrated trajectories using LLM-driven reflection, enabling agents to learn from improved decision-making processes. We finally leverage these calibrated trajectories with successful trajectories for reinforced training. Extensive experiments demonstrate that STeCa significantly outperforms existing methods. Further analysis highlights that timely calibration enables agents to complete tasks with greater robustness. Our code and data are available at https://github.com/WangHanLinHenry/STeCa.
Authors: Hojoon Lee, Youngdo Lee, Takuma Seno, Donghu Kim, Peter Stone, Jaegul Choo
Abstract: Scaling up the model size and computation has brought consistent performance improvements in supervised learning. However, this lesson often fails to apply to reinforcement learning (RL) because training the model on non-stationary data easily leads to overfitting and unstable optimization. In response, we introduce SimbaV2, a novel RL architecture designed to stabilize optimization by (i) constraining the growth of weight and feature norm by hyperspherical normalization; and (ii) using a distributional value estimation with reward scaling to maintain stable gradients under varying reward magnitudes. Using the soft actor-critic as a base algorithm, SimbaV2 scales up effectively with larger models and greater compute, achieving state-of-the-art performance on 57 continuous control tasks across 4 domains. The code is available at https://dojeon-ai.github.io/SimbaV2.
Authors: Yanyang Li, Michael Lyu, Liwei Wang
Abstract: Solving complex tasks in a single attempt is challenging for large language models (LLMs). Iterative interaction with the environment and feedback is often required to achieve success, making effective feedback utilization a critical topic. Existing approaches either struggle with length generalization or rely on naive retries without leveraging prior information. In this paper, we introduce FTTT, a novel paradigm that formulates feedback utilization as an optimization problem at test time. Additionally, we propose a learnable test-time optimizer, OpTune, to effectively exploit feedback. Experiments on two LLMs across four reasoning datasets demonstrate that FTTT and OpTune achieve superior scalability and performance.
Authors: Zhipeng Cheng, Xiaoyu Xia, Hong Wang, Minghui Liwang, Ning Chen, Xuwei Fan, Xianbin Wang
Abstract: Edge inference (EI) has emerged as a promising paradigm to address the growing limitations of cloud-based Deep Neural Network (DNN) inference services, such as high response latency, limited scalability, and severe data privacy exposure. However, deploying DNN models on resource-constrained edge devices introduces additional challenges, including limited computation/storage resources, dynamic service demands, and heightened privacy risks. To tackle these issues, this paper presents a novel privacy-aware optimization framework that jointly addresses DNN model deployment, user-server association, and model partitioning, with the goal of minimizing long-term average inference delay under resource and privacy constraints. The problem is formulated as a complex, NP-hard stochastic optimization. To efficiently handle system dynamics and computational complexity, we employ a Lyapunov-based approach to transform the long-term objective into tractable per-slot decisions. Furthermore, we introduce a coalition formation game to enable adaptive user-server association and design a greedy algorithm for model deployment within each coalition. Extensive simulations demonstrate that the proposed algorithm significantly reduces inference delay and consistently satisfies privacy constraints, outperforming state-of-the-art baselines across diverse scenarios.
Authors: Ruifeng Tan, Weixiang Hong, Jiayue Tang, Xibin Lu, Ruijun Ma, Xiang Zheng, Jia Li, Jiaqiang Huang, Tong-Yi Zhang
Abstract: Battery Life Prediction (BLP), which relies on time series data produced by battery degradation tests, is crucial for battery utilization, optimization, and production. Despite impressive advancements, this research area faces three key challenges. Firstly, the limited size of existing datasets impedes insights into modern battery life data. Secondly, most datasets are restricted to small-capacity lithium-ion batteries tested under a narrow range of diversity in labs, raising concerns about the generalizability of findings. Thirdly, inconsistent and limited benchmarks across studies obscure the effectiveness of baselines and leave it unclear if models popular in other time series fields are effective for BLP. To address these challenges, we propose BatteryLife, a comprehensive dataset and benchmark for BLP. BatteryLife integrates 16 datasets, offering a 2.5 times sample size compared to the previous largest dataset, and provides the most diverse battery life resource with batteries from 8 formats, 59 chemical systems, 9 operating temperatures, and 421 charge/discharge protocols, including both laboratory and industrial tests. Notably, BatteryLife is the first to release battery life datasets of zinc-ion batteries, sodium-ion batteries, and industry-tested large-capacity lithium-ion batteries. With the comprehensive dataset, we revisit the effectiveness of baselines popular in this and other time series fields. Furthermore, we propose CyclePatch, a plug-in technique that can be employed in various neural networks. Extensive benchmarking of 18 methods reveals that models popular in other time series fields can be unsuitable for BLP, and CyclePatch consistently improves model performance establishing state-of-the-art benchmarks. Moreover, BatteryLife evaluates model performance across aging conditions and domains. BatteryLife is available at https://github.com/Ruifeng-Tan/BatteryLife.
Authors: Tobias Kreiman, Aditi S. Krishnapriyan
Abstract: Machine Learning Force Fields (MLFFs) are a promising alternative to expensive ab initio quantum mechanical molecular simulations. Given the diversity of chemical spaces that are of interest and the cost of generating new data, it is important to understand how MLFFs generalize beyond their training distributions. In order to characterize and better understand distribution shifts in MLFFs, we conduct diagnostic experiments on chemical datasets, revealing common shifts that pose significant challenges, even for large foundation models trained on extensive data. Based on these observations, we hypothesize that current supervised training methods inadequately regularize MLFFs, resulting in overfitting and learning poor representations of out-of-distribution systems. We then propose two new methods as initial steps for mitigating distribution shifts for MLFFs. Our methods focus on test-time refinement strategies that incur minimal computational cost and do not use expensive ab initio reference labels. The first strategy, based on spectral graph theory, modifies the edges of test graphs to align with graph structures seen during training. Our second strategy improves representations for out-of-distribution systems at test-time by taking gradient steps using an auxiliary objective, such as a cheap physical prior. Our test-time refinement strategies significantly reduce errors on out-of-distribution systems, suggesting that MLFFs are capable of and can move towards modeling diverse chemical spaces, but are not being effectively trained to do so. Our experiments establish clear benchmarks for evaluating the generalization capabilities of the next generation of MLFFs. Our code is available at https://tkreiman.github.io/projects/mlff_distribution_shifts/.
URLs: https://tkreiman.github.io/projects/mlff_distribution_shifts/.
Authors: Subhajit Maity, Killian Hitsman, Xin Li, Aritra Dutta
Abstract: Kolmogorov-Arnold networks (KANs) are a remarkable innovation consisting of learnable activation functions with the potential to capture more complex relationships from data. Presently, KANs are deployed by replacing multilayer perceptrons (MLPs) in deep networks, including advanced architectures such as vision Transformers (ViTs). This work asks whether a similar replacement in the attention can bring benefits. In this paper, we design the first learnable attention called Kolmogorov-Arnold Attention (KArAt) for ViTs that can operate on any basis, ranging from Fourier, Wavelets, Splines, to Rational Functions. However, learnable activations in attention cause a memory explosion. To remedy this, we propose a modular version of KArAt that uses a low-rank approximation. By adopting the Fourier basis, Fourier-KArAt and its variants, in some cases, outperform their traditional softmax counterparts, or show comparable performance on CIFAR-10, CIFAR-100, and ImageNet-1K datasets. We also deploy Fourier KArAt to ConViT and Swin-Transformer, and use it in detection and segmentation with ViT-Det. We dissect these architectures' performance by analyzing their loss landscapes, weight distributions, optimizer path, attention visualization, and transferability to other datasets. KArAt's learnable activation shows a better attention score across all ViTs, indicating better token-to-token interactions, contributing to better inference. Still, its generalizability does not scale with larger ViTs. However, many factors, including the present computing interface, affect the performance of parameter- and memory-heavy KArAts. We note that the goal of this paper is not to produce efficient attention or challenge the traditional activations; by designing KArAt, we are the first to show that attention can be learned and encourage researchers to explore KArAt in conjunction with more advanced architectures.
Authors: Jasmeet Kaur
Abstract: Effective uncertainty quantification is important for training modern predictive models with limited data, enhancing both accuracy and robustness. While Bayesian methods are effective for this purpose, they can be challenging to scale. When employing approximate Bayesian inference, ensuring the quality of samples from the posterior distribution in a computationally efficient manner is essential. This paper addresses the estimation of the Bayesian posterior to generate diverse samples by approximating the gradient flow of the Kullback-Leibler (KL) divergence and the cross entropy of the target approximation under the metric induced by the Stein Operator. It presents empirical evaluations on classification tasks to assess the method's performance and discuss its effectiveness for Model-Based Reinforcement Learning that uses uncertainty-aware network dynamics models.
Authors: Tao Feng, Yihang Sun, Jiaxuan You
Abstract: The powerful capabilities of Large Language Models (LLMs) have led to their growing use in evaluating human-generated content, particularly in evaluating research ideas within academic settings. Existing solutions primarily rely on prompt-based LLM methods or fine-tuned lightweight language models for idea evaluation. However, these methods are often unstable and struggle to comprehend the complex semantic information embedded in the ideas, impeding their ability to perform high-quality evaluations. To address the above challenges, we propose GraphEval, a lightweight graph-based LLM framework for idea evaluation. Our insight is that a complex idea can be broken down into comprehensible viewpoint nodes using prompts from small LLMs. These viewpoint nodes can then be linked together through edges created from LLM-based relation extraction and/or BERT similarity scores. The created viewpoint-graph can be used to conveniently propagate scores across view-nodes to improve the robustness of the idea evaluations. In particular, we propose two lightweight graph-based methods for idea evaluation: (1) GraphEval-LP: a training-free label propagation algorithm that propagates evaluation scores from known view-nodes to unknown nodes; (2) GraphEval-GNN: a Graph Neural Networks (GNN) that is trained to predict the evaluation scores given the observed graph with minimal computation resources. Moreover, to overcome LLM's limitation in objectively assessing the novelty of ideas, we further propose a novelty detection model to GraphEval-GNN to enhance its capability in judging idea novelty. Experiments on two datasets show GraphEval improves F1 scores by at least 14% with low computation and API costs. Additionally, GraphEval can effectively detect plagiarized ideas.
Authors: Ricardo N. Ferreira, Cl\'audia Soares
Abstract: Constrained Online Convex Optimization (COCO) can be seen as a generalization of the standard Online Convex Optimization (OCO) framework. At each round, a cost function and constraint function are revealed after a learner chooses an action. The goal is to minimize both the regret and cumulative constraint violation (CCV) against an adaptive adversary. We show for the first time that is possible to obtain the optimal $O(\sqrt{T})$ bound on both regret and CCV, improving the best known bounds of $O \left( \sqrt{T} \right)$ and $\tilde{O} \left( \sqrt{T} \right)$ for the regret and CCV, respectively. Based on a new surrogate loss function enforcing a minimum penalty on the constraint function, we demonstrate that both the Follow-the-Regularized-Leader and the Online Gradient Descent achieve the optimal bounds.
Authors: Chuan Qin, Xin Chen, Chengrui Wang, Pengmin Wu, Xi Chen, Yihang Cheng, Jingyi Zhao, Meng Xiao, Xiangchao Dong, Qingqing Long, Boya Pan, Han Wu, Chengzan Li, Yuanchun Zhou, Hui Xiong, Hengshu Zhu
Abstract: In recent years, the rapid advancement of Artificial Intelligence (AI) technologies, particularly Large Language Models (LLMs), has revolutionized the paradigm of scientific discovery, establishing AI-for-Science (AI4Science) as a dynamic and evolving field. However, there is still a lack of an effective framework for the overall assessment of AI4Science, particularly from a holistic perspective on data quality and model capability. Therefore, in this study, we propose SciHorizon, a comprehensive assessment framework designed to benchmark the readiness of AI4Science from both scientific data and LLM perspectives. First, we introduce a generalizable framework for assessing AI-ready scientific data, encompassing four key dimensions: Quality, FAIRness, Explainability, and Compliance-which are subdivided into 15 sub-dimensions. Drawing on data resource papers published between 2018 and 2023 in peer-reviewed journals, we present recommendation lists of AI-ready datasets for Earth, Life, and Materials Sciences, making a novel and original contribution to the field. Concurrently, to assess the capabilities of LLMs across multiple scientific disciplines, we establish 16 assessment dimensions based on five core indicators Knowledge, Understanding, Reasoning, Multimodality, and Values spanning Mathematics, Physics, Chemistry, Life Sciences, and Earth and Space Sciences. Using the developed benchmark datasets, we have conducted a comprehensive evaluation of over 50 representative open-source and closed source LLMs. All the results are publicly available and can be accessed online at www.scihorizon.cn/en.
Authors: Jiexia Ye, Weiqi Zhang, Ziyue Li, Jia Li, Fugee Tsung
Abstract: Medical time series (MedTS) classification is crucial for improved diagnosis in healthcare, and yet it is challenging due to the varying granularity of patterns, intricate inter-channel correlation, information redundancy, and label scarcity. While existing transformer-based models have shown promise in time series analysis, they mainly focus on forecasting and fail to fully exploit the distinctive characteristics of MedTS data. In this paper, we introduce Sparseformer, a transformer specifically designed for MedTS classification. We propose a sparse token-based dual-attention mechanism that enables global modeling and token compression, allowing dynamic focus on the most informative tokens while distilling redundant features. This mechanism is then applied to the multi-granularity, cross-channel encoding of medical signals, capturing intra- and inter-granularity correlations and inter-channel connections. The sparsification design allows our model to handle heterogeneous inputs of varying lengths and channels directly. Further, we introduce an adaptive label encoder to address label space misalignment across datasets, equipping our model with cross-dataset transferability to alleviate the medical label scarcity issue. Our model outperforms 12 baselines across seven medical datasets under supervised learning. In the few-shot learning experiments, our model also achieves superior average results. In addition, the in-domain and cross-domain experiments among three diagnostic scenarios demonstrate our model's zero-shot learning capability. Collectively, these findings underscore the robustness and transferability of our model in various medical applications.
Authors: Jihwan Oh, Minchan Jeong, Jongwoo Ko, Se-Young Yun
Abstract: Large Language Models $($LLMs$)$ solve complex problems using training-free methods like prompt engineering and in-context learning, yet ensuring reasoning correctness remains challenging. While self-correction methods such as self-consistency and self-refinement aim to improve reliability, they often reinforce biases due to the lack of effective feedback mechanisms. Multi-Agent Debate $($MAD$)$ has emerged as an alternative, but we identify two key limitations: bias reinforcement, where debate amplifies model biases instead of correcting them, and lack of perspective diversity, as all agents share the same model and reasoning patterns, limiting true debate effectiveness. To systematically evaluate these issues, we introduce $\textit{MetaNIM Arena}$, a benchmark designed to assess LLMs in adversarial strategic decision-making, where dynamic interactions influence optimal decisions. To overcome MAD's limitations, we propose $\textbf{DReaMAD}$ $($$\textbf{D}$iverse $\textbf{Rea}$soning via $\textbf{M}$ulti-$\textbf{A}$gent $\textbf{D}$ebate with Refined Prompt$)$, a novel framework that $(1)$ refines LLM's strategic prior knowledge to improve reasoning quality and $(2)$ promotes diverse viewpoints within a single model by systematically modifying prompts, reducing bias. Empirical results show that $\textbf{DReaMAD}$ significantly improves decision accuracy, reasoning diversity, and bias mitigation across multiple strategic tasks, establishing it as a more effective approach for LLM-based decision-making.
Authors: Christoforos N. Spartalis, Theodoros Semertzidis, Efstratios Gavves, Petros Daras
Abstract: We present LoTUS, a novel Machine Unlearning (MU) method that eliminates the influence of training samples from pre-trained models, avoiding retraining from scratch. LoTUS smooths the prediction probabilities of the model up to an information-theoretic bound, mitigating its over-confidence stemming from data memorization. We evaluate LoTUS on Transformer and ResNet18 models against eight baselines across five public datasets. Beyond established MU benchmarks, we evaluate unlearning on ImageNet1k, a large-scale dataset, where retraining is impractical, simulating real-world conditions. Moreover, we introduce the novel Retrain-Free Jensen-Shannon Divergence (RF-JSD) metric to enable evaluation under real-world conditions. The experimental results show that LoTUS outperforms state-of-the-art methods in terms of both efficiency and effectiveness. Code: https://github.com/cspartalis/LoTUS.
Authors: Olawale Salaudeen, Nicole Chiou, Shiny Weng, Sanmi Koyejo
Abstract: Spurious correlations are unstable statistical associations that hinder robust decision-making. Conventional wisdom suggests that models relying on such correlations will fail to generalize out-of-distribution (OOD), especially under strong distribution shifts. However, empirical evidence challenges this view as naive in-distribution empirical risk minimizers often achieve the best OOD accuracy across popular OOD generalization benchmarks. In light of these results, we propose a different perspective: many widely used benchmarks for evaluating robustness to spurious correlations are misspecified. Specifically, they fail to include shifts in spurious correlations that meaningfully impact OOD generalization, making them unsuitable for evaluating the benefit of removing such correlations. We establish conditions under which a distribution shift can reliably assess a model's reliance on spurious correlations. Crucially, under these conditions, we should not observe a strong positive correlation between in-distribution and OOD accuracy, often called "accuracy on the line." Yet, most state-of-the-art benchmarks exhibit this pattern, suggesting they do not effectively assess robustness. Our findings expose a key limitation in current benchmarks used to evaluate domain generalization algorithms, that is, models designed to avoid spurious correlations. We highlight the need to rethink how robustness to spurious correlations is assessed, identify well-specified benchmarks the field should prioritize, and enumerate strategies for designing future benchmarks that meaningfully reflect robustness under distribution shift.
Authors: Xinyu Luo, Kecheng Chen, Pao-Sheng Vincent Sun, Chris Xing Tian, Arindam Basu, Haoliang Li
Abstract: Spiking Neural Networks (SNNs), as a biologically plausible alternative to Artificial Neural Networks (ANNs), have demonstrated advantages in terms of energy efficiency, temporal processing, and biological plausibility. However, SNNs are highly sensitive to distribution shifts, which can significantly degrade their performance in real-world scenarios. Traditional test-time adaptation (TTA) methods designed for ANNs often fail to address the unique computational dynamics of SNNs, such as sparsity and temporal spiking behavior. To address these challenges, we propose SPike-Aware Consistency Enhancement (SPACE), the first source-free and single-instance TTA method specifically designed for SNNs. SPACE leverages the inherent spike dynamics of SNNs to maximize the consistency of spike-behavior-based local feature maps across augmented versions of a single test sample, enabling robust adaptation without requiring source data. We evaluate SPACE on multiple datasets. Furthermore, SPACE exhibits robust generalization across diverse network architectures, consistently enhancing the performance of SNNs on CNNs (such as VGG and ResNet), Transformer models, and ConvLSTM architectures. Experimental results show that SPACE outperforms state-of-the-art methods, highlighting its effectiveness and robustness in real-world settings.
Authors: Mei Li, Yuxiang Lu, Qinyan Dai, Suizhi Huang, Yue Ding, Hongtao Lu
Abstract: Continual Learning (CL) strives to learn incrementally across tasks while mitigating catastrophic forgetting. A key challenge in CL is balancing stability (retaining prior knowledge) and plasticity (learning new tasks). While representative gradient projection methods ensure stability, they often limit plasticity. Model merging techniques offer promising solutions, but prior methods typically rely on empirical assumptions and carefully selected hyperparameters. In this paper, we explore the potential of model merging to enhance the stability-plasticity trade-off, providing theoretical insights that underscore its benefits. Specifically, we reformulate the merging mechanism using Bayesian continual learning principles and derive a closed-form solution for the optimal merging coefficient that adapts to the diverse characteristics of tasks. To validate our approach, we introduce a two-stage framework named BECAME, which synergizes the expertise of gradient projection and adaptive merging. Extensive experiments show that our approach outperforms state-of-the-art CL methods and existing merging strategies.
Authors: Mete Erdogan, Cengiz Pehlevan, Alper T. Erdogan
Abstract: We introduce Error Broadcast and Decorrelation (EBD), a novel learning framework for neural networks that addresses credit assignment by directly broadcasting output errors to individual layers, circumventing weight transport of backpropagation. EBD is rigorously grounded in the stochastic orthogonality property of Minimum Mean Square Error estimators. This fundamental principle states that the error of an optimal estimator is orthogonal to functions of the input. Guided by this insight, EBD defines layerwise loss functions that directly penalize correlations between layer activations and output errors, thereby establishing a principled foundation for error broadcasting. This theoretically sound mechanism naturally leads to the experimentally observed three-factor learning rule and integrates with biologically plausible frameworks to enhance performance and plausibility. Numerical experiments demonstrate EBD's competitive or better performance against other error-broadcast methods on benchmark datasets. Our findings establish EBD as an efficient, biologically plausible, and principled alternative for neural network training.
Authors: Sahil Rajesh Dhayalkar
Abstract: We propose a combinatorial and graph-theoretic theory of dropout by modeling training as a random walk over a high-dimensional graph of binary subnetworks. Each node represents a masked version of the network, and dropout induces stochastic traversal across this space. We define a subnetwork contribution score that quantifies generalization and show that it varies smoothly over the graph. Using tools from spectral graph theory, PAC-Bayes analysis, and combinatorics, we prove that generalizing subnetworks form large, connected, low-resistance clusters, and that their number grows exponentially with network width. This reveals dropout as a mechanism for sampling from a robust, structured ensemble of well-generalizing subnetworks with built-in redundancy. Extensive experiments validate every theoretical claim across diverse architectures. Together, our results offer a unified foundation for understanding dropout and suggest new directions for mask-guided regularization and subnetwork optimization.
Authors: Jianhao Yan, Yafu Li, Zican Hu, Zhi Wang, Ganqu Cui, Xiaoye Qu, Yu Cheng, Yue Zhang
Abstract: Recent advances in large reasoning models (LRMs) demonstrate that sophisticated behaviors such as multi-step reasoning and self-reflection can emerge via reinforcement learning with verifiable rewards~(\textit{RLVR}). However, existing \textit{RLVR} approaches are inherently ``on-policy'', limiting learning to a model's own outputs and failing to acquire reasoning abilities beyond its initial capabilities. To address this issue, we introduce \textbf{LUFFY} (\textbf{L}earning to reason \textbf{U}nder o\textbf{FF}-polic\textbf{Y} guidance), a framework that augments \textit{RLVR} with off-policy reasoning traces. LUFFY dynamically balances imitation and exploration by combining off-policy demonstrations with on-policy rollouts during training. Specifically, LUFFY combines the Mixed-Policy GRPO framework, which has a theoretically guaranteed convergence rate, alongside policy shaping via regularized importance sampling to avoid superficial and rigid imitation during mixed-policy training. Compared with previous RLVR methods, LUFFY achieves an over \textbf{+6.4} average gain across six math benchmarks and an advantage of over \textbf{+6.2} points in out-of-distribution tasks. Most significantly, we show that LUFFY successfully trains weak models in scenarios where on-policy RLVR completely fails. These results provide compelling evidence that LUFFY transcends the fundamental limitations of on-policy RLVR and demonstrates the great potential of utilizing off-policy guidance in RLVR.
Authors: Ruiquan Huang, Yingbin Liang, Jing Yang
Abstract: Language recognition tasks are fundamental in natural language processing (NLP) and have been widely used to benchmark the performance of large language models (LLMs). These tasks also play a crucial role in explaining the working mechanisms of transformers. In this work, we focus on two representative tasks in the category of regular language recognition, known as `even pairs' and `parity check', the aim of which is to determine whether the occurrences of certain subsequences in a given sequence are even. Our goal is to explore how a one-layer transformer, consisting of an attention layer followed by a linear layer, learns to solve these tasks by theoretically analyzing its training dynamics under gradient descent. While even pairs can be solved directly by a one-layer transformer, parity check need to be solved by integrating Chain-of-Thought (CoT), either into the inference stage of a transformer well-trained for the even pairs task, or into the training of a one-layer transformer. For both problems, our analysis shows that the joint training of attention and linear layers exhibits two distinct phases. In the first phase, the attention layer grows rapidly, mapping data sequences into separable vectors. In the second phase, the attention layer becomes stable, while the linear layer grows logarithmically and approaches in direction to a max-margin hyperplane that correctly separates the attention layer outputs into positive and negative samples, and the loss decreases at a rate of $O(1/t)$. Our experiments validate those theoretical results.
Authors: Xinyue Zeng, Haohui Wang, Junhong Lin, Jun Wu, Tyler Cody, Dawei Zhou
Abstract: The proliferation of open-sourced Large Language Models (LLMs) and diverse downstream tasks necessitates efficient model selection, given the impracticality of fine-tuning all candidates due to computational constraints. Despite the recent advances in LLM selection, a fundamental research question largely remains nascent: how can we model the dynamic behaviors of LLMs during fine-tuning, thereby enhancing our understanding of their generalization performance across diverse downstream tasks? In this work, we propose a novel theoretical framework that provides a proper lens to assess the generalization capabilities of LLMs, thereby enabling accurate and efficient LLM selection for downstream applications. In particular, we first derive a PAC-Bayesian Generalization Bound that unveils fine-tuning dynamics of LLMs and then introduce LENSLLM, a Neural Tangent Kernel (NTK)-based Rectified Scaling Model that enables accurate performance predictions across diverse tasks while maintaining computational efficiency. Extensive empirical results on 3 large-scale benchmarks demonstrate that our model achieves up to 91.1% accuracy and reduces up to 88.5% computational cost in LLM selection, outperforming 5 state-of-the-art methods. We open-source our proposed LENSLLM model and corresponding results at LensLLM.io.
Authors: Wei Chen, Shigui Li, Jiacheng Li, Junmei Yang, John Paisley, Delu Zeng
Abstract: Density ratio estimation is fundamental to tasks involving f-divergences, yet existing methods often fail under significantly different distributions or inadequately overlapping supports -- the density-chasm and the support-chasm problems. Additionally, prior approaches yield divergent time scores near boundaries, leading to instability. We design $\textbf{D}^3\textbf{RE}$, a unified framework for robust, stable and efficient density ratio estimation. We propose the dequantified diffusion bridge interpolant (DDBI), which expands support coverage and stabilizes time scores via diffusion bridges and Gaussian dequantization. Building on DDBI, the proposed dequantified Schr{\"o}dinger bridge interpolant (DSBI) incorporates optimal transport to solve the Schr{\"o}dinger bridge problem, enhancing accuracy and efficiency. Our method offers uniform approximation and bounded time scores in theory, and outperforms baselines empirically in mutual information and density estimation tasks.
Authors: Steven Song, Morgan Borjigin-Wang, Irene Madejski, Robert L. Grossman
Abstract: The Cancer Genome Atlas (TCGA) has enabled novel discoveries and served as a large-scale reference through its harmonized genomics, clinical, and image data. Prior studies have trained bespoke cancer survival prediction models from unimodal or multimodal TCGA data. A modern paradigm in biomedical deep learning is the development of foundation models (FMs) to derive meaningful feature embeddings, agnostic to a specific modeling task. Biomedical text especially has seen growing development of FMs. While TCGA contains free-text data as pathology reports, these have been historically underutilized. Here, we investigate the feasibility of training classical, multimodal survival models over zero-shot embeddings extracted by FMs. We show the ease and additive effect of multimodal fusion, outperforming unimodal models. We demonstrate the benefit of including pathology report text and rigorously evaluate the effect of model-based text summarization and hallucination. Overall, we modernize survival modeling by leveraging FMs and information extraction from pathology reports.
Authors: David Grasev
Abstract: Gas turbine engines represent complex highly nonlinear dynamical systems. Deriving their physics-based models can be challenging as it requires performance characteristics, that are not always available, and one often has to make many simplifying assumptions. In this paper, the limitations of conventional experimental methods used to derive component-level and locally linear parameter-varying models are discussed and addressed by employing identification techniques based on data collected from standard engine operation under closed-loop control. The rotor dynamics were estimated using the sparse identification of nonlinear dynamics. Subsequently, the autonomous part of the dynamics was mapped into an optimally constructed Koopman eigenfunction space. The process included eigenvalue optimization using metaheuristic algorithms and temporal projection, followed by gradient-based eigenfunction identification. The resulting Koopman model was validated against an in-house reference component-level model. A globally optimal nonlinear feedback controller and a Kalman estimator were then designed in the eigenfunction space and compared to the classical and gain-scheduled proportional-integral controllers, as well as a proposed internal model control approach. The eigenmode structure allowed targeting individual modes during the optimization process, resulting in a better performance tuning. The results showed that the Koopman-based controller outperformed the other benchmark controllers in both reference tracking and disturbance rejection, under sea-level and varying flight conditions, due to its global nature.
Authors: Wesley A Suttle, Vipul K Sharma, Brian M Sadler
Abstract: Multi-agent reinforcement learning (MARL) methods typically require that agents enjoy global state observability, preventing development of decentralized algorithms and limiting scalability. Recent work has shown that, under assumptions on decaying inter-agent influence, global observability can be replaced by local neighborhood observability at each agent, enabling decentralization and scalability. Real-world applications enjoying such decay properties remain underexplored, however, despite the fact that signal power decay, or signal attenuation, due to path loss is an intrinsic feature of many problems in wireless communications and radar networks. In this paper, we show that signal attenuation enables decentralization in MARL by considering the illustrative special case of performing power allocation for target detection in a radar network. To achieve this, we propose two new constrained multi-agent Markov decision process formulations of this power allocation problem, derive local neighborhood approximations for global value function and policy gradient estimates and establish corresponding error bounds, and develop decentralized saddle point policy gradient algorithms for solving the proposed problems. Our approach, though oriented towards the specific radar network problem we consider, provides a useful model for extensions to additional problems in wireless communications and radar networks.
Authors: Sahil Rajesh Dhayalkar
Abstract: We develop a novel theoretical framework for analyzing ReLU neural networks through the lens of a combinatorial object we term the ReLU Transition Graph (RTG). In this graph, each node corresponds to a linear region induced by the network's activation patterns, and edges connect regions that differ by a single neuron flip. Building on this structure, we derive a suite of new theoretical results connecting RTG geometry to expressivity, generalization, and robustness. Our contributions include tight combinatorial bounds on RTG size and diameter, a proof of RTG connectivity, and graph-theoretic interpretations of VC-dimension. We also relate entropy and average degree of the RTG to generalization error. Each theoretical result is rigorously validated via carefully controlled experiments across varied network depths, widths, and data regimes. This work provides the first unified treatment of ReLU network structure via graph theory and opens new avenues for compression, regularization, and complexity control rooted in RTG analysis.
Authors: Sahil Rajesh Dhayalkar
Abstract: We present a complete theoretical and empirical framework establishing feedforward neural networks as universal finite-state machines (N-FSMs). Our results prove that finite-depth ReLU and threshold networks can exactly simulate deterministic finite automata (DFAs) by unrolling state transitions into depth-wise neural layers, with formal characterizations of required depth, width, and state compression. We demonstrate that DFA transitions are linearly separable, binary threshold activations allow exponential compression, and Myhill-Nerode equivalence classes can be embedded into continuous latent spaces while preserving separability. We also formalize the expressivity boundary: fixed-depth feedforward networks cannot recognize non-regular languages requiring unbounded memory. Unlike prior heuristic or probing-based studies, we provide constructive proofs and design explicit DFA-unrolled neural architectures that empirically validate every claim. Our results bridge deep learning, automata theory, and neural-symbolic computation, offering a rigorous blueprint for how discrete symbolic processes can be realized in continuous neural systems.
Authors: Nikolaos-Lysias Kosioris, Sotirios Nikoletseas, Gavrilis Filios, Stefanos Panagiotou
Abstract: The rapid increase in computing power and the ability to store Big Data in the infrastructure has enabled predictions in a large variety of domains by Machine Learning. However, in many cases, existing Machine Learning tools are considered insufficient or incorrect since they exploit only probabilistic dependencies rather than inference logic. Causal Machine Learning methods seem to close this gap. In this paper, two prevalent tools based on Causal Machine Learning methods are compared, as well as their mathematical underpinning background. The operation of the tools is demonstrated by examining their response to 18 queries, based on the IDEAL Household Energy Dataset, published by the University of Edinburgh. First, it was important to evaluate the causal relations assumption that allowed the use of this approach; this was based on the preexisting scientific knowledge of the domain and was implemented by use of the in-built validation tools. Results were encouraging and may easily be extended to other domains.
Authors: Shuo Han, German Espinosa, Junda Huang, Daniel A. Dombeck, Malcolm A. MacIver, Bradly C. Stadie
Abstract: Recent advances in reinforcement learning (RL) have demonstrated impressive capabilities in complex decision-making tasks. This progress raises a natural question: how do these artificial systems compare to biological agents, which have been shaped by millions of years of evolution? To help answer this question, we undertake a comparative study of biological mice and RL agents in a predator-avoidance maze environment. Through this analysis, we identify a striking disparity: RL agents consistently demonstrate a lack of self-preservation instinct, readily risking ``death'' for marginal efficiency gains. These risk-taking strategies are in contrast to biological agents, which exhibit sophisticated risk-assessment and avoidance behaviors. Towards bridging this gap between the biological and artificial, we propose two novel mechanisms that encourage more naturalistic risk-avoidance behaviors in RL agents. Our approach leads to the emergence of naturalistic behaviors, including strategic environment assessment, cautious path planning, and predator avoidance patterns that closely mirror those observed in biological systems.
Authors: Conor Rowan, Alireza Doostan
Abstract: Though neural networks trained on large datasets have been successfully used to describe and predict many physical phenomena, there is a sense among scientists that, unlike traditional scientific models comprising simple mathematical expressions, their findings cannot be integrated into the body of scientific knowledge. Critics of machine learning's inability to produce human-understandable relationships have converged on the concept of "interpretability" as its point of departure from more traditional forms of science. As the growing interest in interpretability has shown, researchers in the physical sciences seek not just predictive models, but also to uncover the fundamental principles that govern a system of interest. However, clarity around a definition of interpretability and the precise role that it plays in science is lacking in the literature. In this work, we argue that researchers in equation discovery and symbolic regression tend to conflate the concept of sparsity with interpretability. We review key papers on interpretable machine learning from outside the scientific community and argue that, though the definitions and methods they propose can inform questions of interpretability for scientific machine learning (SciML), they are inadequate for this new purpose. Noting these deficiencies, we propose an operational definition of interpretability for the physical sciences. Our notion of interpretability emphasizes understanding of the mechanism over mathematical sparsity. Innocuous though it may seem, this emphasis on mechanism shows that sparsity is often unnecessary. It also questions the possibility of interpretable scientific discovery when prior knowledge is lacking. We believe a precise and philosophically informed definition of interpretability in SciML will help focus research efforts toward the most significant obstacles to realizing a data-driven scientific future.
Authors: Roberto L. Castro, Andrei Panferov, Soroush Tabesh, Oliver Sieberling, Jiale Chen, Mahdi Nikdan, Saleh Ashkboos, Dan Alistarh
Abstract: Training large language models (LLMs) models directly in low-precision offers a way to address computational costs by improving both throughput and energy efficiency. For those purposes, NVIDIA's recent Blackwell architecture facilitates very low-precision operations using FP4 variants. Yet, current algorithms for training LLMs in FP4 precision face significant accuracy degradation and often rely on mixed-precision fallbacks. In this paper, we investigate hardware-supported FP4 training and introduce a new approach for accurate, end-to-end FP4 training with all the major computations (i.e., linear layers) in low precision. Through extensive evaluations on Llama-type models, we reveal a new low-precision scaling law that quantifies performance trade-offs across bit-widths and training setups. Guided by this investigation, we design an "optimal" technique in terms of accuracy-vs-computation, called Quartet. We implement Quartet using optimized CUDA kernels tailored for Blackwell, demonstrating that fully FP4-based training is a competitive alternative to FP16 half-precision and to FP8 training. Our code is available at https://github.com/IST-DASLab/Quartet.
Authors: Huazi Pan, Yanjun Zhang, Leo Yu Zhang, Scott Adams, Abbas Kouzani, Suiyang Khoo
Abstract: Manipulation of local training data and local updates, i.e., the poisoning attack, is the main threat arising from the collaborative nature of the federated learning (FL) paradigm. Most existing poisoning attacks aim to manipulate local data/models in a way that causes denial-of-service (DoS) issues. In this paper, we introduce a novel attack method, named Federated Learning Sliding Attack (FedSA) scheme, aiming at precisely introducing the extent of poisoning in a subtle controlled manner. It operates with a predefined objective, such as reducing global model's prediction accuracy by 10%. FedSA integrates robust nonlinear control-Sliding Mode Control (SMC) theory with model poisoning attacks. It can manipulate the updates from malicious clients to drive the global model towards a compromised state, achieving this at a controlled and inconspicuous rate. Additionally, leveraging the robust control properties of FedSA allows precise control over the convergence bounds, enabling the attacker to set the global accuracy of the poisoned model to any desired level. Experimental results demonstrate that FedSA can accurately achieve a predefined global accuracy with fewer malicious clients while maintaining a high level of stealth and adjustable learning rates.
Authors: Rohan Ghuge, Vidya Muthukumar, Sahil Singla
Abstract: We study \emph{online multicalibration}, a framework for ensuring calibrated predictions across multiple groups in adversarial settings, across $T$ rounds. Although online calibration is typically studied in the $\ell_1$ norm, prior approaches to online multicalibration have taken the indirect approach of obtaining rates in other norms (such as $\ell_2$ and $\ell_{\infty}$) and then transferred these guarantees to $\ell_1$ at additional loss. In contrast, we propose a direct method that achieves improved and oracle-efficient rates of $\widetilde{\mathcal{O}}(T^{-1/3})$ and $\widetilde{\mathcal{O}}(T^{-1/4})$ respectively, for online $\ell_1$-multicalibration. Our key insight is a novel reduction of online \(\ell_1\)-multicalibration to an online learning problem with product-based rewards, which we refer to as \emph{online linear-product optimization} ($\mathtt{OLPO}$). To obtain the improved rate of $\widetilde{\mathcal{O}}(T^{-1/3})$, we introduce a linearization of $\mathtt{OLPO}$ and design a no-regret algorithm for this linearized problem. Although this method guarantees the desired sublinear rate (nearly matching the best rate for online calibration), it is computationally expensive when the group family \(\mathcal{H}\) is large or infinite, since it enumerates all possible groups. To address scalability, we propose a second approach to $\mathtt{OLPO}$ that makes only a polynomial number of calls to an offline optimization (\emph{multicalibration evaluation}) oracle, resulting in \emph{oracle-efficient} online \(\ell_1\)-multicalibration with a rate of $\widetilde{\mathcal{O}}(T^{-1/4})$. Our framework also extends to certain infinite families of groups (e.g., all linear functions on the context space) by exploiting a $1$-Lipschitz property of the \(\ell_1\)-multicalibration error with respect to \(\mathcal{H}\).
Authors: Deyang Kong, Qi Guo, Xiangyu Xi, Wei Wang, Jingang Wang, Xunliang Cai, Shikun Zhang, Wei Ye
Abstract: Reinforcement learning exhibits potential in enhancing the reasoning abilities of large language models, yet it is hard to scale for the low sample efficiency during the rollout phase. Existing methods attempt to improve efficiency by scheduling problems based on problem difficulties. However, these approaches suffer from unstable and biased estimations of problem difficulty and fail to capture the alignment between model competence and problem difficulty in RL training, leading to suboptimal results. To tackle these limitations, this paper introduces $\textbf{C}$ompetence-$\textbf{D}$ifficulty $\textbf{A}$lignment $\textbf{S}$ampling ($\textbf{CDAS}$), which enables accurate and stable estimation of problem difficulties by aggregating historical performance discrepancies of problems. Then the model competence is quantified to adaptively select problems whose difficulty is in alignment with the model's current competence using a fixed-point system. Experimental results across a range of challenging mathematical benchmarks show that CDAS achieves great improvements in both accuracy and efficiency. CDAS attains the highest average accuracy against baselines and exhibits significant speed advantages compared to Dynamic Sampling, a competitive strategy in DAPO, which is 2.33 times slower than CDAS.
Authors: Chi Zhang, Ziying Jia, George K. Atia, Sihong He, Yue Wang
Abstract: Transfer reinforcement learning aims to derive a near-optimal policy for a target environment with limited data by leveraging abundant data from related source domains. However, it faces two key challenges: the lack of performance guarantees for the transferred policy, which can lead to undesired actions, and the risk of negative transfer when multiple source domains are involved. We propose a novel framework based on the pessimism principle, which constructs and optimizes a conservative estimation of the target domain's performance. Our framework effectively addresses the two challenges by providing an optimized lower bound on target performance, ensuring safe and reliable decisions, and by exhibiting monotonic improvement with respect to the quality of the source domains, thereby avoiding negative transfer. We construct two types of conservative estimations, rigorously characterize their effectiveness, and develop efficient distributed algorithms with convergence guarantees. Our framework provides a theoretically sound and practically robust solution for transfer learning in reinforcement learning.
Authors: Bingdong Li, Mei Jiang, Hong Qian, Ke Tang, Aimin Zhou, Peng Yang
Abstract: Evolutionary Reinforcement Learning (ERL), training the Reinforcement Learning (RL) policies with Evolutionary Algorithms (EAs), have demonstrated enhanced exploration capabilities and greater robustness than using traditional policy gradient. However, ERL suffers from the high computational costs and low search efficiency, as EAs require evaluating numerous candidate policies with expensive simulations, many of which are ineffective and do not contribute meaningfully to the training. One intuitive way to reduce the ineffective evaluations is to adopt the surrogates. Unfortunately, existing ERL policies are often modeled as deep neural networks (DNNs) and thus naturally represented as high-dimensional vectors containing millions of weights, which makes the building of effective surrogates for ERL policies extremely challenging. This paper proposes a novel surrogate-assisted ERL that integrates Autoencoders (AE) and Hyperbolic Neural Networks (HNN). Specifically, AE compresses high-dimensional policies into low-dimensional representations while extracting key features as the inputs for the surrogate. HNN, functioning as a classification-based surrogate model, can learn complex nonlinear relationships from sampled data and enable more accurate pre-selection of the sampled policies without real evaluations. The experiments on 10 Atari and 4 Mujoco games have verified that the proposed method outperforms previous approaches significantly. The search trajectories guided by AE and HNN are also visually demonstrated to be more effective, in terms of both exploration and convergence. This paper not only presents the first learnable policy embedding and surrogate-modeling modules for high-dimensional ERL policies, but also empirically reveals when and why they can be successful.
Authors: Juwei Yue, Haikuo Li, Jiawei Sheng, Yihan Guo, Xinghua Zhang, Chuan Zhou, Tingwen Liu, Li Guo
Abstract: Dynamics modeling has been introduced as a novel paradigm in message passing (MP) of graph neural networks (GNNs). Existing methods consider MP between nodes as a heat diffusion process, and leverage heat equation to model the temporal evolution of nodes in the embedding space. However, heat equation can hardly depict the wave nature of graph signals in graph signal processing. Besides, heat equation is essentially a partial differential equation (PDE) involving a first partial derivative of time, whose numerical solution usually has low stability, and leads to inefficient model training. In this paper, we would like to depict more wave details in MP, since graph signals are essentially wave signals that can be seen as a superposition of a series of waves in the form of eigenvector. This motivates us to consider MP as a wave propagation process to capture the temporal evolution of wave signals in the space. Based on wave equation in physics, we innovatively develop a graph wave equation to leverage the wave propagation on graphs. In details, we demonstrate that the graph wave equation can be connected to traditional spectral GNNs, facilitating the design of graph wave networks based on various Laplacians and enhancing the performance of the spectral GNNs. Besides, the graph wave equation is particularly a PDE involving a second partial derivative of time, which has stronger stability on graphs than the heat equation that involves a first partial derivative of time. Additionally, we theoretically prove that the numerical solution derived from the graph wave equation are constantly stable, enabling to significantly enhance model efficiency while ensuring its performance. Extensive experiments show that GWNs achieve SOTA and efficient performance on benchmark datasets, and exhibit outstanding performance in addressing challenging graph problems, such as over-smoothing and heterophily.
Authors: Lijun Zhang, Lin Li, Yajie Qi, Huizhong Song, Yaodong Yang, Jun Wang, Wei Wei
Abstract: When fine-tuning pre-trained Large Language Models (LLMs) to align with human values and intentions, maximizing the estimated reward can lead to superior performance, but it also introduces potential risks due to deviations from the reference model's intended behavior. Most existing methods typically introduce KL divergence to constrain deviations between the trained model and the reference model; however, this may not be sufficient in certain applications that require tight risk control. In this paper, we introduce Risk-aware Direct Preference Optimization (Ra-DPO), a novel approach that incorporates risk-awareness by employing a class of nested risk measures. This approach formulates a constrained risk-aware advantage function maximization problem and then converts the Bradley-Terry model into a token-level representation. The objective function maximizes the likelihood of the policy while suppressing the deviation between a trained model and the reference model using a sequential risk ratio, thereby enhancing the model's risk-awareness. Experimental results across three open-source datasets: IMDb Dataset, Anthropic HH Dataset, and AlpacaEval, demonstrate the proposed method's superior performance in balancing alignment performance and model drift. Our code is opensourced at https://github.com/zlj123-max/Ra-DPO.
Authors: Dane Malenfant, Blake A. Richards
Abstract: Sometimes we benefit from actions that others have taken even when we are unaware that they took those actions. For example, if your neighbor chooses not to take a parking spot in front of your house when you are not there, you can benefit, even without being aware that they took this action. These "hidden gifts" represent an interesting challenge for multi-agent reinforcement learning (MARL), since assigning credit when the beneficial actions of others are hidden is non-trivial. Here, we study the impact of hidden gifts with a very simple MARL task. In this task, agents in a grid-world environment have individual doors to unlock in order to obtain individual rewards. As well, if all the agents unlock their door the group receives a larger collective reward. However, there is only one key for all of the doors, such that the collective reward can only be obtained when the agents drop the key for others after they use it. Notably, there is nothing to indicate to an agent that the other agents have dropped the key, thus the act of dropping the key for others is a "hidden gift". We show that several different state-of-the-art RL algorithms, including MARL algorithms, fail to learn how to obtain the collective reward in this simple task. Interestingly, we find that independent model-free policy gradient agents can solve the task when we provide them with information about their own action history, but MARL agents still cannot solve the task with action history. Finally, we derive a correction term for these independent agents, inspired by learning aware approaches, which reduces the variance in learning and helps them to converge to collective success more reliably. These results show that credit assignment in multi-agent settings can be particularly challenging in the presence of "hidden gifts", and demonstrate that learning awareness in independent agents can benefit these settings.
Authors: Anissa Alloula, Charles Jones, Ben Glocker, Bart{\l}omiej W. Papie\.z
Abstract: Despite the constant development of new bias mitigation methods for machine learning, no method consistently succeeds, and a fundamental question remains unanswered: when and why do bias mitigation techniques fail? In this paper, we hypothesise that a key factor may be the often-overlooked but crucial step shared by many bias mitigation methods: the definition of subgroups. To investigate this, we conduct a comprehensive evaluation of state-of-the-art bias mitigation methods across multiple vision and language classification tasks, systematically varying subgroup definitions, including coarse, fine-grained, intersectional, and noisy subgroups. Our results reveal that subgroup choice significantly impacts performance, with certain groupings paradoxically leading to worse outcomes than no mitigation at all. Our findings suggest that observing a disparity between a set of subgroups is not a sufficient reason to use those subgroups for mitigation. Through theoretical analysis, we explain these phenomena and uncover a counter-intuitive insight that, in some cases, improving fairness with respect to a particular set of subgroups is best achieved by using a different set of subgroups for mitigation. Our work highlights the importance of careful subgroup definition in bias mitigation and presents it as an alternative lever for improving the robustness and fairness of machine learning models.
Authors: M\'onika Farsang, Ramin Hasani, Radu Grosu
Abstract: We present LrcSSM, a $\textit{nonlinear}$ recurrent model that processes long sequences as fast as today's linear state-space layers. By forcing the state-transition matrix to be diagonal and learned at every step, the full sequence can be solved in parallel with a single prefix-scan, giving $\mathcal{O}(TD)$ time and memory and only $\mathcal{O}(\log T)$ sequential depth, for input-sequence length $T$ and a state dimension $D$. Moreover, LrcSSM offers a formal gradient-stability guarantee that other input-varying systems such as Liquid-S4 and Mamba do not provide. Lastly, for network depth $L$, as the forward and backward passes cost $\Theta(T\,D\,L)$ FLOPs, with its low sequential depth and parameter count $\Theta(D\,L)$, the model follows the compute-optimal scaling law regime ($\beta \approx 0.42$) recently observed for Mamba, outperforming quadratic-attention Transformers at equal compute while avoiding the memory overhead of FFT-based long convolutions. We show that on a series of long-range forecasting tasks, LrcSSM outperforms LRU, S5 and Mamba.
Authors: Tommy Xu, Zhitian Zhang, Xiangyu Sun, Lauren Kelly Zung, Hossein Hajimirsadeghi, Greg Mori
Abstract: Predictive modeling on tabular data is the cornerstone of many real-world applications. Although gradient boosting machines and some recent deep models achieve strong performance on tabular data, they often lack interpretability. On the other hand, large language models (LLMs) have demonstrated powerful capabilities to generate human-like reasoning and explanations, but remain under-performed for tabular data prediction. In this paper, we propose a new approach that leverages reasoning-based LLMs, trained using reinforcement learning, to perform more accurate and explainable predictions on tabular data. Our method introduces custom reward functions that guide the model not only toward high prediction accuracy but also toward human-understandable reasons for its predictions. Experimental results show that our model achieves promising performance on financial benchmark datasets, outperforming most existing LLMs.
Authors: Ruijie Li, Xiang Zhao, Qiao Ning, Shikai Guo
Abstract: In tennis tournaments, momentum, a critical yet elusive phenomenon, reflects the dynamic shifts in performance of athletes that can decisively influence match outcomes. Despite its significance, momentum in terms of effective modeling and multi-granularity analysis across points, games, sets, and matches in tennis tournaments remains underexplored. In this study, we define a novel Momentum Score (MS) metric to quantify a player's momentum level in multi-granularity tennis tournaments, and design HydraNet, a momentum-driven state-space duality-based framework, to model MS by integrating thirty-two heterogeneous dimensions of athletes performance in serve, return, psychology and fatigue. HydraNet integrates a Hydra module, which builds upon a state-space duality (SSD) framework, capturing explicit momentum with a sliding-window mechanism and implicit momentum through cross-game state propagation. It also introduces a novel Versus Learning method to better enhance the adversarial nature of momentum between the two athletes at a macro level, along with a Collaborative-Adversarial Attention Mechanism (CAAM) for capturing and integrating intra-player and inter-player dynamic momentum at a micro level. Additionally, we construct a million-level tennis cross-tournament dataset spanning from 2012-2023 Wimbledon and 2013-2023 US Open, and validate the multi-granularity modeling capability of HydraNet for the MS metric on this dataset. Extensive experimental evaluations demonstrate that the MS metric constructed by the HydraNet framework provides actionable insights into how momentum impacts outcomes at different granularities, establishing a new foundation for momentum modeling and sports analysis. To the best of our knowledge, this is the first work to explore and effectively model momentum across multiple granularities in professional tennis tournaments.
Authors: Yu-Heng Hung, Kai-Jie Lin, Yu-Heng Lin, Chien-Yi Wang, Cheng Sun, Ping-Chun Hsieh
Abstract: Bayesian optimization (BO) offers an efficient pipeline for optimizing black-box functions with the help of a Gaussian process prior and an acquisition function (AF). Recently, in the context of single-objective BO, learning-based AFs witnessed promising empirical results given its favorable non-myopic nature. Despite this, the direct extension of these approaches to multi-objective Bayesian optimization (MOBO) suffer from the \textit{hypervolume identifiability issue}, which results from the non-Markovian nature of MOBO problems. To tackle this, inspired by the non-Markovian RL literature and the success of Transformers in language modeling, we present a generalized deep Q-learning framework and propose \textit{BOFormer}, which substantiates this framework for MOBO via sequence modeling. Through extensive evaluation, we demonstrate that BOFormer constantly outperforms the benchmark rule-based and learning-based algorithms in various synthetic MOBO and real-world multi-objective hyperparameter optimization problems. We have made the source code publicly available to encourage further research in this direction.
Authors: Jujie He, Jiacai Liu, Chris Yuhao Liu, Rui Yan, Chaojie Wang, Peng Cheng, Xiaoyu Zhang, Fuxiang Zhang, Jiacheng Xu, Wei Shen, Siyuan Li, Liang Zeng, Tianwen Wei, Cheng Cheng, Bo An, Yang Liu, Yahui Zhou
Abstract: The success of DeepSeek-R1 underscores the significant role of reinforcement learning (RL) in enhancing the reasoning capabilities of large language models (LLMs). In this work, we present Skywork-OR1, an effective and scalable RL implementation for long Chain-of-Thought (CoT) models. Building on the DeepSeek-R1-Distill model series, our RL approach achieves notable performance gains, increasing average accuracy across AIME24, AIME25, and LiveCodeBench from 57.8% to 72.8% (+15.0%) for the 32B model and from 43.6% to 57.5% (+13.9%) for the 7B model. Our Skywork-OR1-32B model surpasses both DeepSeek-R1 and Qwen3-32B on the AIME24 and AIME25 benchmarks, while achieving comparable results on LiveCodeBench. The Skywork-OR1-7B and Skywork-OR1-Math-7B models demonstrate competitive reasoning capabilities among models of similar size. We perform comprehensive ablation studies on the core components of our training pipeline to validate their effectiveness. Additionally, we thoroughly investigate the phenomenon of entropy collapse, identify key factors affecting entropy dynamics, and demonstrate that mitigating premature entropy collapse is critical for improved test performance. To support community research, we fully open-source our model weights, training code, and training datasets.
Authors: Haomiao Qiu, Miao Zhang, Ziyue Qiao, Weili Guan, Min Zhang, Liqiang Nie
Abstract: Continual Learning requires a model to learn multiple tasks in sequence while maintaining both stability:preserving knowledge from previously learned tasks, and plasticity:effectively learning new tasks. Gradient projection has emerged as an effective and popular paradigm in CL, where it partitions the gradient space of previously learned tasks into two orthogonal subspaces: a primary subspace and a minor subspace. New tasks are learned effectively within the minor subspace, thereby reducing interference with previously acquired knowledge. However, existing Gradient Projection methods struggle to achieve an optimal balance between plasticity and stability, as it is hard to appropriately partition the gradient space. In this work, we consider a continual learning paradigm based on Low-Rank Adaptation, which has gained considerable attention due to its efficiency and wide applicability, and propose a novel approach for continual learning, called SplitLoRA. We first provide a theoretical analysis of how subspace partitioning affects model stability and plasticity. Informed by this analysis, we then introduce an effective method that derives the optimal partition of the gradient space for previously learned tasks. This approach effectively balances stability and plasticity in continual learning. Experimental results on multiple datasets demonstrate that the proposed method achieves state-of-the-art performance.
Authors: Haomiao Qiu, Miao Zhang, Ziyue Qiao, Liqiang Nie
Abstract: Continual Learning (CL) aims to enable models to continuously acquire new knowledge from a sequence of tasks with avoiding the forgetting of learned information. However, existing CL methods only rely on the parameters of the most recent task for inference, which makes them susceptible to catastrophic forgetting. Inspired by the recent success of model merging techniques, we propose \textbf{Perturb-and-Merge (P\&M)}, a novel continual learning framework that integrates model merging into the CL paradigm to mitigate forgetting. Specifically, after training on each task, P\&M constructs a new model by forming a convex combination of the previous model and the newly trained task-specific model. Through theoretical analysis, we minimize the total loss increase across all tasks and derive an analytical solution for the optimal merging coefficient. To further improve the performance of the merged model, we observe that the degradation introduced during merging can be alleviated by a regularization term composed of the task vector and the Hessian matrix of the loss function. Interestingly, we show that this term can be efficiently approximated using second-order symmetric finite differences, and a stochastic perturbation strategy along the task vector direction is accordingly devised which incurs no additional forward or backward passes while providing an effective approximation of the regularization term. Finally, we combine P\&M with LoRA, a parameter-efficient fine-tuning method, to reduce memory overhead. Our proposed approach achieves state-of-the-art performance on several continual learning benchmark datasets.
Authors: Mihir Prabhudesai, Lili Chen, Alex Ippoliti, Katerina Fragkiadaki, Hao Liu, Deepak Pathak
Abstract: Reinforcement learning (RL) has enabled machine learning models to achieve significant advances in many fields. Most recently, RL has empowered frontier language models to solve challenging math, science, and coding problems. However, central to any RL algorithm is the reward function, and reward engineering is a notoriously difficult problem in any domain. In this paper, we propose RENT: Reinforcement Learning via Entropy Minimization -- a fully unsupervised RL method that requires no external reward or ground-truth answers, and instead uses the model's entropy of its underlying distribution as an intrinsic reward. We find that by reinforcing the chains of thought that yield high model confidence on its generated answers, the model improves its reasoning ability. In our experiments, we showcase these improvements on an extensive suite of commonly-used reasoning benchmarks, including GSM8K, MATH500, AMC, AIME, and GPQA, and models of varying sizes from the Qwen and Mistral families. The generality of our unsupervised learning method lends itself to applicability in a wide range of domains where external supervision is unavailable.
Authors: Hyunsik Yoo, Zhichen Zeng, Jian Kang, Ruizhong Qiu, David Zhou, Zhining Liu, Fei Wang, Charlie Xu, Eunice Chan, Hanghang Tong
Abstract: User-side group fairness is crucial for modern recommender systems, aiming to alleviate performance disparities among user groups defined by sensitive attributes like gender, race, or age. In the ever-evolving landscape of user-item interactions, continual adaptation to newly collected data is crucial for recommender systems to stay aligned with the latest user preferences. However, we observe that such continual adaptation often exacerbates performance disparities. This necessitates a thorough investigation into user-side fairness in dynamic recommender systems, an area that has been unexplored in the literature. This problem is challenging due to distribution shifts, frequent model updates, and non-differentiability of ranking metrics. To our knowledge, this paper presents the first principled study on ensuring user-side fairness in dynamic recommender systems. We start with theoretical analyses on fine-tuning v.s. retraining, showing that the best practice is incremental fine-tuning with restart. Guided by our theoretical analyses, we propose FAir Dynamic rEcommender (FADE), an end-to-end fine-tuning framework to dynamically ensure user-side fairness over time. To overcome the non-differentiability of recommendation metrics in the fairness loss, we further introduce Differentiable Hit (DH) as an improvement over the recent NeuralNDCG method, not only alleviating its gradient vanishing issue but also achieving higher efficiency. Besides that, we also address the instability issue of the fairness loss by leveraging the competing nature between the recommendation loss and the fairness loss. Through extensive experiments on real-world datasets, we demonstrate that FADE effectively and efficiently reduces performance disparities with little sacrifice in the overall recommendation performance.
Authors: Zelai Xu, Chao Yu, Fei Fang, Yu Wang, Yi Wu
Abstract: Agents built with large language models (LLMs) have shown great potential across a wide range of domains. However, in complex decision-making tasks, pure LLM-based agents tend to exhibit intrinsic bias in their choice of actions, which is inherited from the model's training data and results in suboptimal performance. To develop strategic language agents, i.e., agents that generate flexible language actions and possess strong decision-making abilities, we propose a novel framework that powers LLM-based agents with reinforcement learning (RL). We consider Werewolf, a popular social deduction game, as a challenging testbed that emphasizes versatile communication and strategic gameplay. To mitigate the intrinsic bias in language actions, our agents use an LLM to perform deductive reasoning and generate a diverse set of action candidates. Then an RL policy trained to optimize the decision-making ability chooses an action from the candidates to play in the game. Extensive experiments show that our agents overcome the intrinsic bias and outperform existing LLM-based agents in the Werewolf game. We also conduct human-agent experiments and find that our agents achieve human-level performance and demonstrate strong strategic play.
Authors: Akihiro Fujii, Anh Khoa Augustin Lu, Koji Shimizu, Satoshi Watanabe
Abstract: Materials design aims to discover novel compounds with desired properties. However, prevailing strategies face critical trade-offs. Conventional element-substitution approaches readily and adaptively incorporate various domain knowledge but remain confined to a narrow search space. In contrast, deep generative models efficiently explore vast compositional landscapes, yet they struggle to flexibly integrate domain knowledge. To address these trade-offs, we propose a gradient-based material design framework that combines these strengths, offering both efficiency and adaptability. In our method, chemical compositions are optimised to achieve target properties by using property prediction models and their gradients. In order to seamlessly enforce diverse constraints, including those reflecting domain insights such as oxidation states, discretised compositional ratios, types of elements, and their abundance, we apply masks and employ a special loss function, namely the integer loss. Furthermore, we initialise the optimisation using promising candidates from existing dataset, effectively guiding the search away from unfavourable regions and thus helping to avoid poor solutions. Our approach demonstrates a more efficient exploration of superconductor candidates, uncovering candidate materials with higher critical temperature than conventional element-substitution and generative models. Importantly, it could propose new compositions beyond those found in existing databases, including new hydride superconductors absent from the training dataset but which share compositional similarities with materials found in literature. This synergy of domain knowledge and machine-learning-based scalability provides a robust foundation for rapid, adaptive, and comprehensive materials design for superconductors and beyond.
Authors: Qianglin Wen, Chengchun Shi, Ying Yang, Niansheng Tang, Hongtu Zhu
Abstract: A/B testing has become the gold standard for policy evaluation in modern technological industries. Motivated by the widespread use of switchback experiments in A/B testing, this paper conducts a comprehensive comparative analysis of various switchback designs in Markovian environments. Unlike many existing works which derive the optimal design based on specific and relatively simple estimators, our analysis covers a range of state-of-the-art estimators developed in the reinforcement learning (RL) literature. It reveals that the effectiveness of different switchback designs depends crucially on (i) the size of the carryover effect and (ii) the auto-correlations among reward errors over time. Meanwhile, these findings are estimator-agnostic, i.e., they apply to most RL estimators. Based on these insights, we provide a workflow to offer guidelines for practitioners on designing switchback experiments in A/B testing.
Authors: Jing Lin, Yao Feng, Weiyang Liu, Michael J. Black
Abstract: Numerous methods have been proposed to detect, estimate, and analyze properties of people in images, including 3D pose, shape, contact, human-object interaction, and emotion. While widely applicable in vision and other areas, such methods require expert knowledge to select, use, and interpret the results. To address this, we introduce ChatHuman, a language-driven system that integrates the capabilities of specialized methods into a unified framework. ChatHuman functions as an assistant proficient in utilizing, analyzing, and interacting with tools specific to 3D human tasks, adeptly discussing and resolving related challenges. Built on a Large Language Model (LLM) framework, ChatHuman is trained to autonomously select, apply, and interpret a diverse set of tools in response to user inputs. Our approach overcomes significant hurdles in adapting LLMs to 3D human tasks, including the need for domain-specific knowledge and the ability to interpret complex 3D outputs. The innovations of ChatHuman include leveraging academic publications to instruct the LLM on tool usage, employing a retrieval-augmented generation model to create in-context learning examples for managing new tools, and effectively discriminating between and integrating tool results by transforming specialized 3D outputs into comprehensible formats. Experiments demonstrate that ChatHuman surpasses existing models in both tool selection accuracy and overall performance across various 3D human tasks, and it supports interactive chatting with users. ChatHuman represents a significant step toward consolidating diverse analytical methods into a unified, robust system for 3D human tasks.
Authors: Yeoneung Kim, Gihun Kim, Jiwhan Park, Insoon Yang
Abstract: We propose a novel Thompson sampling algorithm that learns linear quadratic regulators (LQR) with a Bayesian regret bound of $O(\sqrt{T})$. Our method leverages Langevin dynamics with a carefully designed preconditioner and incorporates a simple excitation mechanism. We show that the excitation signal drives the minimum eigenvalue of the preconditioner to grow over time, thereby accelerating the approximate posterior sampling process. Furthermore, we establish nontrivial concentration properties of the approximate posteriors generated by our algorithm. These properties enable us to bound the moments of the system state and attain an $O(\sqrt{T})$ regret bound without relying on the restrictive assumptions that are often used in the literature.
Authors: Eshant English, Eliot Wong-Toi, Matteo Fontana, Stephan Mandt, Padhraic Smyth, Christoph Lippert
Abstract: Conformal prediction provides machine learning models with prediction sets that offer theoretical guarantees, but the underlying assumption of exchangeability limits its applicability to time series data. Furthermore, existing approaches struggle to handle multi-step ahead prediction tasks, where uncertainty estimates across multiple future time points are crucial. We propose JANET (Joint Adaptive predictioN-region Estimation for Time-series), a novel framework for constructing conformal prediction regions that are valid for both univariate and multivariate time series. JANET generalises the inductive conformal framework and efficiently produces joint prediction regions with controlled K-familywise error rates, enabling flexible adaptation to specific application needs. Our empirical evaluation demonstrates JANET's superior performance in multi-step prediction tasks across diverse time series datasets, highlighting its potential for reliable and interpretable uncertainty quantification in sequential data.
Authors: Yupeng Chang, Yi Chang, Yuan Wu
Abstract: Large language models (LLMs) have demonstrated remarkable proficiency across various natural language processing (NLP) tasks. However, adapting LLMs to downstream applications requires computationally intensive and memory-demanding fine-tuning procedures. To alleviate these burdens, parameter-efficient fine-tuning (PEFT) techniques have emerged as a promising approach to tailor LLMs with minimal computational overhead. While PEFT methods offer substantial advantages, they do not fully address the pervasive issue of bias propagation from pre-training data. This work introduces Bias-Alleviating Low-Rank Adaptation (BA-LoRA), a novel PEFT method designed to counteract bias inheritance. BA-LoRA incorporates three distinct regularization terms: (1) a consistency regularizer, (2) a diversity regularizer, and (3) a singular value decomposition regularizer. These regularizers aim to enhance the models' consistency, diversity, and generalization capabilities during fine-tuning. We conduct extensive experiments on natural language understanding (NLU) and natural language generation (NLG) tasks using prominent LLMs such as LLaMA, Mistral, and Gemma. The results demonstrate that BA-LoRA outperforms LoRA and its state-of-the-art variants. Moreover, the extended experiments demonstrate that our method effectively mitigates the adverse effects of pre-training bias, leading to more reliable and robust model outputs. The code is available at https://github.com/cyp-jlu-ai/BA-LoRA.
Authors: Doron Cohen, Aryeh Kontorovich, Roi Weiss
Abstract: We revisit the recently introduced Local Glivenko-Cantelli setting, which studies distribution-dependent uniform convergence rates of the Empirical Mean Estimator (EME). In this work, we investigate generalizations of this setting where arbitrary estimators are allowed rather than just the EME. Can a strictly larger class of measures be learned? Can better risk decay rates be obtained? We provide exhaustive answers to these questions, which are both negative, provided the learner is barred from exploiting some infinite-dimensional pathologies. On the other hand, allowing such exploits does lead to a strictly larger class of learnable measures.
Authors: Zeyu Zhang, Sixu Yan, Muzhi Han, Zaijin Wang, Xinggang Wang, Song-Chun Zhu, Hangxin Liu
Abstract: We propose M3Bench, a new benchmark for whole-body motion generation in mobile manipulation tasks. Given a 3D scene context, M3Bench requires an embodied agent to reason about its configuration, environmental constraints, and task objectives to generate coordinated whole-body motion trajectories for object rearrangement. M3Bench features 30,000 object rearrangement tasks across 119 diverse scenes, providing expert demonstrations generated by our newly developed M3BenchMaker, an automatic data generation tool that produces whole-body motion trajectories from high-level task instructions using only basic scene and robot information. Our benchmark includes various task splits to evaluate generalization across different dimensions and leverages realistic physics simulation for trajectory assessment. Extensive evaluation analysis reveals that state-of-the-art models struggle with coordinating base-arm motion while adhering to environmental and task-specific constraints, underscoring the need for new models to bridge this gap. By releasing M3Bench and M3BenchMaker we aim to advance robotics research toward more adaptive and capable mobile manipulation in diverse, real-world environments.
Authors: Akihiro Fujii, Yoshitaka Ushiku, Koji Shimizu, Anh Khoa Augustin Lu, Satoshi Watanabe
Abstract: Gradient-based methods offer a simple, efficient strategy for materials design by directly optimizing candidates using gradients from pretrained property predictors. However, their use in crystal structure optimization is hindered by two key challenges: handling non-differentiable constraints, such as charge neutrality and structural fidelity, and susceptibility to poor local minima. We revisit and extend the gradient-based methods to address these issues. We propose Simultaneous Multi-property Optimization using Adaptive Crystal Synthesizer (SMOACS), which integrates oxidation-number masks and template-based initialization to enforce non-differentiable constraints, avoid poor local minima, and flexibly incorporate additional constraints without retraining. SMOACS enables multi-property optimization. including exceptional targets such as high-temperature superconductivity, and scales to large crystal systems, both persistent challenges for generative models, even those enhanced with gradient-based guidance from property predictors. In experiments on five target properties and three datasets, SMOACS outperforms generative models and Bayesian optimization methods, successfully designing 135-atom perovskite structures that satisfy multiple property targets and constraints, a task at which the other methods fail entirely.
Authors: Haozhen Zhang, Tao Feng, Jiaxuan You
Abstract: Retrieval-augmented generation (RAG) has revitalized Large Language Models (LLMs) by injecting non-parametric factual knowledge. Compared with long-context LLMs, RAG is considered an effective summarization tool in a more concise and lightweight manner, which can interact with LLMs multiple times using diverse queries to get comprehensive responses. However, the LLM-generated historical responses, which contain potentially insightful information, are largely neglected and discarded by existing approaches, leading to suboptimal results. In this paper, we propose $\textit{graph of records}$ ($\textbf{GoR}$), which leverages historical responses generated by LLMs to enhance RAG for long-context global summarization. Inspired by the $\textit{retrieve-then-generate}$ paradigm of RAG, we construct a graph by establishing an edge between the retrieved text chunks and the corresponding LLM-generated response. To further uncover the intricate correlations between them, GoR features a $\textit{graph neural network}$ and an elaborately designed $\textit{BERTScore}$-based objective for self-supervised model training, enabling seamless supervision signal backpropagation between reference summaries and node embeddings. We comprehensively compare GoR with 12 baselines across four long-context summarization datasets, and the results indicate that our proposed method reaches the best performance ($\textit{e.g.}$, 15%, 8%, and 19% improvement over retrievers w.r.t. Rouge-L, Rouge-1, and Rouge-2 on the WCEP dataset). Extensive experiments further demonstrate the effectiveness of GoR.
Authors: Yuchen Jiao, Gen Li
Abstract: Score-based diffusion models have demonstrated outstanding empirical performance in machine learning and artificial intelligence, particularly in generating high-quality new samples from complex probability distributions. Improving the theoretical understanding of diffusion models, with a particular focus on the convergence analysis, has attracted significant attention. In this work, we develop a convergence rate that is adaptive to the smoothness of different target distributions, referred to as instance-dependent bound. Specifically, we establish an iteration complexity of $\min\{d,d^{2/3}L^{1/3},d^{1/3}L\}\varepsilon^{-2/3}$ (up to logarithmic factors), where $d$ denotes the data dimension, and $\varepsilon$ quantifies the output accuracy in terms of total variation (TV) distance. In addition, $L$ represents a relaxed Lipschitz constant, which, in the case of Gaussian mixture models, scales only logarithmically with the number of components, the dimension and iteration number, demonstrating broad applicability.
Authors: Paris Giampouras, HanQin Cai, Rene Vidal
Abstract: In this paper, we focus on a matrix factorization-based approach to recover low-rank {\it asymmetric} matrices from corrupted measurements. We propose an {\it Overparameterized Preconditioned Subgradient Algorithm (OPSA)} and provide, for the first time in the literature, linear convergence rates independent of the rank of the sought asymmetric matrix in the presence of gross corruptions. Our work goes beyond existing results in preconditioned-type approaches addressing their current limitation, i.e., the lack of convergence guarantees in the case of {\it asymmetric matrices of unknown rank}. By applying our approach to (robust) matrix sensing, we highlight its merits when the measurement operator satisfies a mixed-norm restricted isometry property. Lastly, we present extensive numerical experiments that validate our theoretical results and demonstrate the effectiveness of our approach for different levels of overparameterization and outlier corruptions.
Authors: Jiantao Wu, Sara Atito, Zhenhua Feng, Shentong Mo, Josef Kitler, Muhammad Awais
Abstract: The training methods in AI do involve semantically distinct pairs of samples. However, their role typically is to enhance the between class separability. The actual notion of similarity is normally learned from semantically identical pairs. This paper presents SimLAP: a simple framework for learning visual representation from arbitrary pairs. SimLAP explores the possibility of learning similarity from semantically distinct sample pairs. The approach is motivated by the observation that for any pair of classes there exists a subspace in which semantically distinct samples exhibit similarity. This phenomenon can be exploited for a novel method of learning, which optimises the similarity of an arbitrary pair of samples, while simultaneously learning the enabling subspace. The feasibility of the approach will be demonstrated experimentally and its merits discussed.
Authors: Ernst R\"oell, Bastian Rieck
Abstract: Point-cloud synthesis, i.e. the generation of novel point clouds from an input distribution, remains a challenging task, for which numerous complex machine-learning models have been devised. We develop a novel method that encodes geometrical-topological characteristics of point clouds using inner products, leading to a highly-efficient point cloud representation with provable expressivity properties. Integrated into deep learning models, our encoding exhibits high quality in typical tasks like reconstruction, generation, and interpolation, with inference times orders of magnitude faster than existing methods.
Authors: Peiqi Liu, Zhanqiu Guo, Mohit Warke, Soumith Chintala, Chris Paxton, Nur Muhammad Mahi Shafiullah, Lerrel Pinto
Abstract: Significant progress has been made in open-vocabulary mobile manipulation, where the goal is for a robot to perform tasks in any environment given a natural language description. However, most current systems assume a static environment, which limits the system's applicability in real-world scenarios where environments frequently change due to human intervention or the robot's own actions. In this work, we present DynaMem, a new approach to open-world mobile manipulation that uses a dynamic spatio-semantic memory to represent a robot's environment. DynaMem constructs a 3D data structure to maintain a dynamic memory of point clouds, and answers open-vocabulary object localization queries using multimodal LLMs or open-vocabulary features generated by state-of-the-art vision-language models. Powered by DynaMem, our robots can explore novel environments, search for objects not found in memory, and continuously update the memory as objects move, appear, or disappear in the scene. We run extensive experiments on the Stretch SE3 robots in three real and nine offline scenes, and achieve an average pick-and-drop success rate of 70% on non-stationary objects, which is more than a 2x improvement over state-of-the-art static systems. Our code as well as our experiment and deployment videos are open sourced and can be found on our project website: https://dynamem.github.io/
Authors: Thuan Pham, Xingpeng Li
Abstract: Power system networks are often modeled as homogeneous graphs, which limits the ability of graph neural network (GNN) to capture individual generator features at the same nodes. By introducing the proposed virtual node-splitting strategy, generator-level attributes like costs, limits, and ramp rates can be fully captured by GNN models, improving GNN's learning capacity and prediction accuracy. Optimal power flow (OPF) problem is used for real-time grid operations. Limited timeframe motivates studies to create size-reduced OPF (ROPF) models to relieve the computational complexity. In this paper, with virtual node-splitting, a novel two-stage adaptive hierarchical GNN is developed to (i) predict critical lines that would be congested, and then (ii) predict base generators that would operate at the maximum capacity. This will substantially reduce the constraints and variables needed for OPF, creating the proposed ROPFLG model with reduced monitor lines and reduced generator-specific variables and constraints. Two ROPF models, ROPFL and ROPFG, with just reduced lines or generators respectively, are also implemented as additional benchmark models. Case studies show that the proposed ROPFLG consistently outperforms the benchmark full OPF (FOPF) and the other two ROPF methods, achieving significant computational time savings while reliably finding optimal solutions.
Authors: Bijoy Ahmed Saiem, MD Sadik Hossain Shanto, Rakib Ahsan, Md Rafi ur Rashid
Abstract: As the integration of the Large Language Models (LLMs) into various applications increases, so does their susceptibility to misuse, raising significant security concerns. Numerous jailbreak attacks have been proposed to assess the security defense of LLMs. Current jailbreak attacks mainly rely on scenario camouflage, prompt obfuscation, prompt optimization, and prompt iterative optimization to conceal malicious prompts. In particular, sequential prompt chains in a single query can lead LLMs to focus on certain prompts while ignoring others, facilitating context manipulation. This paper introduces SequentialBreak, a novel jailbreak attack that exploits this vulnerability. We discuss several scenarios, not limited to examples like Question Bank, Dialog Completion, and Game Environment, where the harmful prompt is embedded within benign ones that can fool LLMs into generating harmful responses. The distinct narrative structures of these scenarios show that SequentialBreak is flexible enough to adapt to various prompt formats beyond those discussed. Extensive experiments demonstrate that SequentialBreak uses only a single query to achieve a substantial gain of attack success rate over existing baselines against both open-source and closed-source models. Through our research, we highlight the urgent need for more robust and resilient safeguards to enhance LLM security and prevent potential misuse. All the result files and website associated with this research are available in this GitHub repository: https://anonymous.4open.science/r/JailBreakAttack-4F3B/.
URLs: https://anonymous.4open.science/r/JailBreakAttack-4F3B/.
Authors: Kieran A. Murphy, Yujing Zhang, Dani S. Bassett
Abstract: Multivariate information theory provides a general and principled framework for understanding how the components of a complex system are connected. Existing analyses are coarse in nature -- built up from characterizations of discrete subsystems -- and can be computationally prohibitive. In this work, we propose to study the continuous space of possible descriptions of a composite system as a window into its organizational structure. A description consists of specific information conveyed about each of the components, and the space of possible descriptions is equivalent to the space of lossy compression schemes of the components. We introduce a machine learning framework to optimize descriptions that extremize key information theoretic quantities used to characterize organization, such as total correlation and O-information. Through case studies on spin systems, sudoku boards, and letter sequences from natural language, we identify extremal descriptions that reveal how system-wide variation emerges from individual components. By integrating machine learning into a fine-grained information theoretic analysis of composite random variables, our framework opens a new avenues for probing the structure of real-world complex systems.
Authors: Othmane Echchabi, Aya Lahlou, Nizar Talty, Josh Malcolm Manto, Ka Leung Lam
Abstract: Clean water and sanitation are essential for health, well-being, and sustainable development, yet significant global disparities persist. Although the United Nations' Sustainable Development Goal (SDG) 6 clearly defines targets for universal access to clean water and sanitation, limitations in data coverage and openness impede accurate tracking of progress in many countries. To bridge these gaps, this study integrates Afrobarometer survey data, satellite imagery from Landsat 8 and Sentinel-2, and advanced deep learning techniques using Meta's self-supervised Distillation with No Labels (DINO) model to develop a modeling framework for evaluating access to piped water and sewage system across diverse African regions. The modeling framework achieved notable accuracy, with over 96% for piped water and 97% for sewage system access classification. When combined with geospatial population data, validation against official statistics from the United Nations Joint Monitoring Program demonstrated high concordance at the national scale (R2 of 0.95 for piped water access and R2 of 0.85 for sewage system access). The national-level estimates can represent SDG Indicators 6.1.1 and 6.2.1. This approach provides policymakers and stakeholders with an effective, scalable, and cost-efficient tool to pinpoint underserved areas requiring targeted intervention. The methodology developed herein can be adapted for assessing other infrastructure-related SDGs, promoting enhanced monitoring and informed decision-making towards achieving global sustainability objectives.
Authors: An T. Le, Kay Hansel, Jo\~ao Carvalho, Joe Watson, Julen Urain, Armin Biess, Georgia Chalvatzaki, Jan Peters
Abstract: Batch planning is increasingly necessary to quickly produce diverse and quality motion plans for downstream learning applications, such as distillation and imitation learning. This paper presents Global Tensor Motion Planning (GTMP) -- a sampling-based motion planning algorithm comprising only tensor operations. We introduce a novel discretization structure represented as a random multipartite graph, enabling efficient vectorized sampling, collision checking, and search. We provide a theoretical investigation showing that GTMP exhibits probabilistic completeness while supporting modern GPU/TPU. Additionally, by incorporating smooth structures into the multipartite graph, GTMP directly plans smooth splines without requiring gradient-based optimization. Experiments on lidar-scanned occupancy maps and the MotionBenchMarker dataset demonstrate GTMP's computation efficiency in batch planning compared to baselines, underscoring GTMP's potential as a robust, scalable planner for diverse applications and large-scale robot learning tasks.
Authors: Stefan Smeu, Dragos-Alexandru Boldisor, Dan Oneata, Elisabeta Oneata
Abstract: Good datasets are essential for developing and benchmarking any machine learning system. Their importance is even more extreme for safety critical applications such as deepfake detection - the focus of this paper. Here we reveal that two of the most widely used audio-video deepfake datasets suffer from a previously unidentified spurious feature: the leading silence. Fake videos start with a very brief moment of silence and based on this feature alone, we can separate the real and fake samples almost perfectly. As such, previous audio-only and audio-video models exploit the presence of silence in the fake videos and consequently perform worse when the leading silence is removed. To circumvent latching on such unwanted artifact and possibly other unrevealed ones we propose a shift from supervised to unsupervised learning by training models exclusively on real data. We show that by aligning self-supervised audio-video representations we remove the risk of relying on dataset-specific biases and improve robustness in deepfake detection.
Authors: Xiaobao Wu, Liangming Pan, Yuxi Xie, Ruiwen Zhou, Shuai Zhao, Yubo Ma, Mingzhe Du, Rui Mao, Anh Tuan Luu, William Yang Wang
Abstract: Data contamination hinders fair LLM evaluation by introducing test data into newer models' training sets. Existing studies solve this challenge by updating benchmarks with newly collected data. However, they fail to guarantee contamination-free evaluation as the newly collected data may contain pre-existing knowledge, and their benchmark updates rely on intensive human labor. To address these issues, we in this paper propose AntiLeak-Bench, an automated anti-leakage benchmarking framework. Instead of simply using newly collected data, we construct samples with explicitly new knowledge absent from LLMs' training sets, which thus ensures strictly contamination-free evaluation. We further design a fully automated workflow to build and update our benchmark without human labor. This significantly reduces the cost of benchmark maintenance to accommodate emerging LLMs. Through extensive experiments, we highlight that data contamination likely exists before LLMs' cutoff time and demonstrate AntiLeak-Bench effectively overcomes this challenge.
Authors: Wenhui Cui, Haleh Akrami, Anand A. Joshi, Richard M. Leahy
Abstract: Despite the impressive advances achieved using deep learning for functional brain activity analysis, the heterogeneity of functional patterns and the scarcity of imaging data still pose challenges in tasks such as identifying neurological disorders. For functional Magnetic Resonance Imaging (fMRI), while data may be abundantly available from healthy controls, clinical data is often scarce, especially for rare diseases, limiting the ability of models to identify clinically-relevant features. We overcome this limitation by introducing a novel representation learning strategy integrating meta-learning with self-supervised learning to improve the generalization from normal to clinical features. This approach enables generalization to challenging clinical tasks featuring scarce training data. We achieve this by leveraging self-supervised learning on the control dataset to focus on inherent features that are not limited to a particular supervised task and incorporating meta-learning to improve the generalization across domains. To explore the generalizability of the learned representations to unseen clinical applications, we apply the model to four distinct clinical datasets featuring scarce and heterogeneous data for neurological disorder classification. Results demonstrate the superiority of our representation learning strategy on diverse clinically-relevant tasks. Code is publicly available at https://github.com/wenhui0206/MeTSK/tree/main
Authors: Zijian Liu, Zhengyuan Zhou
Abstract: Recently, the study of heavy-tailed noises in first-order nonconvex stochastic optimization has gotten a lot of attention since it was recognized as a more realistic condition as suggested by many empirical observations. Specifically, the stochastic noise (the difference between the stochastic and true gradient) is considered to have only a finite $\mathfrak{p}$-th moment where $\mathfrak{p}\in\left(1,2\right]$ instead of assuming it always satisfies the classical finite variance assumption. To deal with this more challenging setting, people have proposed different algorithms and proved them to converge at an optimal $\mathcal{O}(T^{\frac{1-\mathfrak{p}}{3\mathfrak{p}-2}})$ rate for smooth objectives after $T$ iterations. Notably, all these new-designed algorithms are based on the same technique - gradient clipping. Naturally, one may want to know whether the clipping method is a necessary ingredient and the only way to guarantee convergence under heavy-tailed noises. In this work, by revisiting the existing Batched Normalized Stochastic Gradient Descent with Momentum (Batched NSGDM) algorithm, we provide the first convergence result under heavy-tailed noises but without gradient clipping. Concretely, we prove that Batched NSGDM can achieve the optimal $\mathcal{O}(T^{\frac{1-\mathfrak{p}}{3\mathfrak{p}-2}})$ rate even under the relaxed smooth condition. More interestingly, we also establish the first $\mathcal{O}(T^{\frac{1-\mathfrak{p}}{2\mathfrak{p}}})$ convergence rate in the case where the tail index $\mathfrak{p}$ is unknown in advance, which is arguably the common scenario in practice.
Authors: Muhammad Ahmad Sultan, Chong Chen, Yingmin Liu, Katarzyna Gil, Karolina Zareba, Rizwan Ahmad
Abstract: Objective: To propose and validate an unsupervised MRI reconstruction method that does not require fully sampled k-space data. Materials and Methods: The proposed method, deep image prior with structured sparsity (DISCUS), extends the deep image prior (DIP) by introducing group sparsity to frame-specific code vectors, enabling the discovery of a low-dimensional manifold for capturing temporal variations. \discus was validated using four studies: (I) simulation of a dynamic Shepp-Logan phantom to demonstrate its manifold discovery capabilities, (II) comparison with compressed sensing and DIP-based methods using simulated single-shot late gadolinium enhancement (LGE) image series from six distinct digital cardiac phantoms in terms of normalized mean square error (NMSE) and structural similarity index measure (SSIM), (III) evaluation on retrospectively undersampled single-shot LGE data from eight patients, and (IV) evaluation on prospectively undersampled single-shot LGE data from eight patients, assessed via blind scoring from two expert readers. Results: DISCUS outperformed competing methods, demonstrating superior reconstruction quality in terms of NMSE and SSIM (Studies I--III) and expert reader scoring (Study IV). Discussion: An unsupervised image reconstruction method is presented and validated on simulated and measured data. These developments can benefit applications where acquiring fully sampled data is challenging.
Authors: Soyeong Jeong, Kangsan Kim, Jinheon Baek, Sung Ju Hwang
Abstract: Retrieval-Augmented Generation (RAG) is a powerful strategy for improving the factual accuracy of models by retrieving external knowledge relevant to queries and incorporating it into the generation process. However, existing approaches primarily focus on text, with some recent advancements considering images, and they largely overlook videos, a rich source of multimodal knowledge capable of representing contextual details more effectively than any other modality. While very recent studies explore the use of videos in response generation, they either predefine query-associated videos without retrieval or convert videos into textual descriptions losing multimodal richness. To tackle these, we introduce VideoRAG, a framework that not only dynamically retrieves videos based on their relevance with queries but also utilizes both visual and textual information. The operation of VideoRAG is powered by recent Large Video Language Models (LVLMs), which enable the direct processing of video content to represent it for retrieval and the seamless integration of retrieved videos jointly with queries for response generation. Also, inspired by that the context size of LVLMs may not be sufficient to process all frames in extremely long videos and not all frames are equally important, we introduce a video frame selection mechanism to extract the most informative subset of frames, along with a strategy to extract textual information from videos (as it can aid the understanding of video content) when their subtitles are not available. We experimentally validate the effectiveness of VideoRAG, showcasing that it is superior to relevant baselines. Code is available at https://github.com/starsuzi/VideoRAG.
Authors: Yifan Zhang, Yifeng Liu, Huizhuo Yuan, Zhen Qin, Yang Yuan, Quanquan Gu, Andrew C Yao
Abstract: Scaling language models to handle longer input sequences typically necessitates large key-value (KV) caches, resulting in substantial memory overhead during inference. In this paper, we propose Tensor Product Attention (TPA), a novel attention mechanism that uses tensor decompositions to represent queries, keys, and values compactly, substantially shrinking the KV cache size at inference time. By factorizing these representations into contextual low-rank components and seamlessly integrating with Rotary Position Embedding (RoPE), TPA achieves improved model quality alongside memory efficiency. Based on TPA, we introduce the Tensor Product Attention Transformer,(T6), a new model architecture for sequence modeling. Through extensive empirical evaluation on language modeling tasks, we demonstrate that T6 surpasses or matches the performance of standard Transformer baselines, including Multi-Head Attention (MHA), Multi-Query Attention (MQA), Grouped-Query Attention (GQA), and Multi-Head Latent Attention (MLA) across various metrics, including perplexity and a range of established evaluation benchmarks. Notably, TPA's memory efficiency and computational efficiency at the decoding stage enable processing longer sequences under fixed resource constraints, addressing a critical scalability challenge in modern language models. The code is available at https://github.com/tensorgi/T6.
Authors: Wen Lai, Alexander Fraser, Ivan Titov
Abstract: Parameter-efficient fine-tuning (PEFT) methods, such as LoRA, are commonly used to adapt LLMs. However, the effectiveness of standard PEFT methods is limited in low-resource scenarios with only a few hundred examples. Recent advances in interpretability research have inspired the emergence of activation editing (or steering) techniques, which modify the activations of specific model components. Due to their extremely small parameter counts, these methods show promise for small datasets. However, their performance is highly dependent on identifying the correct modules to edit and often lacks stability across different datasets. In this paper, we propose Joint Localization and Activation Editing (JoLA), a method that jointly learns (1) which heads in the Transformer to edit (2) whether the intervention should be additive, multiplicative, or both and (3) the intervention parameters themselves - the vectors applied as additive offsets or multiplicative scalings to the head output. Through evaluations on three benchmarks spanning commonsense reasoning, natural language understanding, and natural language generation, we demonstrate that JoLA consistently outperforms existing methods. The code for the method is released at https://github.com/wenlai-lavine/jola.
Authors: Jiale Fu, Yuchu Jiang, Junkai Chen, Jiaming Fan, Xin Geng, Xu Yang
Abstract: Large Language Model (LLM) collaborative decoding techniques improve output quality by combining the outputs of multiple models at each generation step, but they incur high computational costs. In this paper, we introduce Collaborative decoding via Speculation (CoS), a novel framework that accelerates collaborative decoding without compromising performance. Inspired by Speculative Decoding--where a small proposal model generates tokens sequentially, and a larger target model verifies them in parallel, our approach builds on two key insights: (1) the verification distribution can be the combined distribution of both the proposal and target models, and (2) alternating each model as the proposer and verifier can further enhance efficiency. We generalize this method to collaboration among n models and theoretically prove that CoS is never slower than standard collaborative decoding, typically achieving faster speed. Extensive experiments demonstrate CoS is 1.11x-2.23x faster than standard collaborative decoding without compromising generation quality. Our code is available at https://github.com/Kamichanw/CoS/.
Authors: Zachary Cooper-Baldock, Stephen Turnock, Karl Sammut
Abstract: Autonomous Underwater Vehicles (AUVs) encounter significant energy, control and navigation challenges in complex underwater environments, particularly during close-proximity operations, such as launch and recovery (LAR), where fluid interactions and wake effects present additional navigational and energy challenges. Traditional path planning methods fail to incorporate these detailed wake structures, resulting in increased energy consumption, reduced control stability, and heightened safety risks. This paper presents a novel wake-informed, 3D path planning approach that fully integrates localized wake effects and global currents into the planning algorithm. Two variants of the A* algorithm - a current-informed planner and a wake-informed planner - are created to assess its validity and two neural network models are then trained to approximate these planners for real-time applications. Both the A* planners and NN models are evaluated using important metrics such as energy expenditure, path length, and encounters with high-velocity and turbulent regions. The results demonstrate a wake-informed A* planner consistently achieves the lowest energy expenditure and minimizes encounters with high-velocity regions, reducing energy consumption by up to 11.3%. The neural network models are observed to offer computational speedup of 6 orders of magnitude, but exhibit 4.51 - 19.79% higher energy expenditures and 9.81 - 24.38% less optimal paths. These findings underscore the importance of incorporating detailed wake structures into traditional path planning algorithms and the benefits of neural network approximations to enhance energy efficiency and operational safety for AUVs in complex 3D domains.
Authors: Yongchao Chen, Yilun Hao, Yueying Liu, Yang Zhang, Chuchu Fan
Abstract: Existing methods fail to effectively steer Large Language Models (LLMs) between textual reasoning and code generation, leaving symbolic computing capabilities underutilized. We introduce CodeSteer, an effective method for guiding LLM code/text generation. We construct a comprehensive benchmark SymBench comprising 37 symbolic tasks with adjustable complexity and also synthesize datasets of 12k multi-turn guidance/generation trajectories and 5.5k guidance comparison pairs. We fine-tune the Llama-3-8B model with a newly designed multi-turn supervised fine-tuning (SFT) and direct preference optimization (DPO). The resulting model, CodeSteerLLM, augmented with the proposed symbolic and self-answer checkers, effectively guides the code/text generation of larger models. Augmenting GPT-4o with CodeSteer raises its average performance score from 53.3 to 86.4, even outperforming the existing best LLM OpenAI o1 (82.7), o1-preview (74.8), and DeepSeek R1 (76.8) across all 37 tasks (28 seen, 9 unseen). Trained for GPT-4o, CodeSteer demonstrates superior generalizability, providing an average 41.8 performance boost on Claude, Mistral, and GPT-3.5. CodeSteer-guided LLMs fully harness symbolic computing to maintain strong performance on highly complex tasks. Models, Datasets, and Codes are available at https://github.com/yongchao98/CodeSteer-v1.0 and https://huggingface.co/yongchao98.
URLs: https://github.com/yongchao98/CodeSteer-v1.0, https://huggingface.co/yongchao98.
Authors: Elliot Meyerson, Xin Qiu
Abstract: Decomposing hard problems into subproblems often makes them easier and more efficient to solve. With large language models (LLMs) crossing critical reliability thresholds for a growing slate of capabilities, there is an increasing effort to decompose systems into sets of LLM-based agents, each of whom can be delegated sub-tasks. However, this decomposition (even when automated) is often intuitive, e.g., based on how a human might assign roles to members of a human team. How close are these role decompositions to optimal? This position paper argues that asymptotic analysis with LLM primitives is needed to reason about the efficiency of such decomposed systems, and that insights from such analysis will unlock opportunities for scaling them. By treating the LLM forward pass as the atomic unit of computational cost, one can separate out the (often opaque) inner workings of a particular LLM from the inherent efficiency of how a set of LLMs are orchestrated to solve hard problems. In other words, if we want to scale the deployment of LLMs to the limit, instead of anthropomorphizing LLMs, asymptotic analysis with LLM primitives should be used to reason about and develop more powerful decompositions of large problems into LLM agents.
Authors: Shicheng Zhou, Zizhou Zhang, Rong Zhang, Yuchen Yin, Chia Hong Chang, Qinyan Shen
Abstract: This paper analyses the investment returns of three stock sectors, Manuf, Hitec, and Other, in the U.S. stock market, based on the Fama-French three-factor model, the Carhart four-factor model, and the Fama-French five-factor model, in order to test the validity of the Fama-French three-factor model, the Carhart four-factor model, and the Fama-French five-factor model for the three sectors of the market. French five-factor model for the three sectors of the market. Also, the LSTM model is used to explore the additional factors affecting stock returns. The empirical results show that the Fama-French five-factor model has better validity for the three segments of the market under study, and the LSTM model has the ability to capture the factors affecting the returns of certain industries, and can better regress and predict the stock returns of the relevant industries. Keywords- Fama-French model; Carhart model; Factor model; LSTM model.
Authors: Getoar Sopa, Juraj Marusic, Marco Avella-Medina, John P. Cunningham
Abstract: In recent years, there has been much work on scaling Bayesian Optimization to high-dimensional problems, for example hyperparameter tuning in large machine learning models. These scalable methods have been successful, finding high objective values much more quickly than traditional global Bayesian Optimization or random search-based methods. At the same time, these large models often use sensitive data, but preservation of Differential Privacy has not scaled alongside these modern Bayesian Optimization procedures. Here we develop a method to privately optimize potentially high-dimensional parameter spaces using privatized Gradient Informative Bayesian Optimization. Our theoretical results show that under suitable conditions, our method converges exponentially fast to a locally optimal parameter configuration, up to a natural privacy error. Moreover, regardless of whether the assumptions are satisfied, we prove that our algorithm maintains privacy and empirically display superior performance to existing methods in the high-dimensional hyperparameter setting.
Authors: Ziyang Guo, Yifan Wu, Jason Hartline, Jessica Hullman
Abstract: Multiple agents -- including humans and AI models -- are increasingly combined to make decisions with the expectation of achieving complementary performance, where the decisions they make together outperform those made individually. However, knowing how to improve the performance of collaborating agents is often difficult without knowing more about what particular information and strategies each agent employs. With a focus on human-AI pairings, we contribute a decision-theoretic framework for characterizing the value of information -- and consequently, opportunities for agents to better exploit available information -- in AI-assisted decision workflows. We present a novel explanation technique (ILIV-SHAP) that adapts SHAP explanations to highlight human-complementing information. We validate the effectiveness of the framework and ILIV-SHAP through a study of human-AI decision-making. We show that our measure of complementary information can be used to identify which AI model will best complement human decisions. We also find that presenting ILIV-SHAP with AI predictions leads to reliably greater reductions in error over non-AI assisted decisions more than vanilla SHAP.
Authors: Jessica Dai, Paula Gradu, Inioluwa Deborah Raji, Benjamin Recht
Abstract: When an individual reports a negative interaction with some system, how can their personal experience be contextualized within broader patterns of system behavior? We study the reporting database problem, where individual reports of adverse events arrive sequentially, and are aggregated over time. In this work, our goal is to identify whether there are subgroups--defined by any combination of relevant features--that are disproportionately likely to experience harmful interactions with the system. We formalize this problem as a sequential hypothesis test, and identify conditions on reporting behavior that are sufficient for making inferences about disparities in true rates of harm across subgroups. We show that algorithms for sequential hypothesis tests can be applied to this problem with a standard multiple testing correction. We then demonstrate our method on real-world datasets, including mortgage decisions and vaccine side effects; on each, our method (re-)identifies subgroups known to experience disproportionate harm using only a fraction of the data that was initially used to discover them.
Authors: Wei Xuan, Meghna Roy Chowdhury, Yi Ding, Yixue Zhao
Abstract: The global mental health crisis is a pressing concern, with college students particularly vulnerable to rising mental health disorders. The widespread use of smartphones among young adults, while offering numerous benefits, has also been linked to negative outcomes such as addiction and regret, significantly impacting well-being. Leveraging the longest longitudinal dataset collected over four college years through passive mobile sensing, this study is the first to examine the relationship between students' smartphone unlocking behaviors and their mental health at scale in real-world settings. We provide the first evidence demonstrating the predictability of phone unlocking behaviors for mental health outcomes based on a large dataset, highlighting the potential of these novel features for future predictive models. Our findings reveal important variations in smartphone usage across genders and locations, offering a deeper understanding of the interplay between digital behaviors and mental health. We highlight future research directions aimed at mitigating adverse effects and promoting digital well-being in this population.
Authors: Yuhui Zhang, Yuchang Su, Chenyu Wang, Tianhong Li, Zoe Wefers, Jeffrey Nirschl, James Burgess, Daisy Ding, Alejandro Lozano, Emma Lundberg, Serena Yeung-Levy
Abstract: Building a virtual cell capable of accurately simulating cellular behaviors in silico has long been a dream in computational biology. We introduce CellFlux, an image-generative model that simulates cellular morphology changes induced by chemical and genetic perturbations using flow matching. Unlike prior methods, CellFlux models distribution-wise transformations from unperturbed to perturbed cell states, effectively distinguishing actual perturbation effects from experimental artifacts such as batch effects -- a major challenge in biological data. Evaluated on chemical (BBBC021), genetic (RxRx1), and combined perturbation (JUMP) datasets, CellFlux generates biologically meaningful cell images that faithfully capture perturbation-specific morphological changes, achieving a 35% improvement in FID scores and a 12% increase in mode-of-action prediction accuracy over existing methods. Additionally, CellFlux enables continuous interpolation between cellular states, providing a potential tool for studying perturbation dynamics. These capabilities mark a significant step toward realizing virtual cell modeling for biomedical research. Project page: https://yuhui-zh15.github.io/CellFlux/.
Authors: Simone Drago, Marco Mussi, Alberto Maria Metelli
Abstract: In this work, we provide a refined analysis of the UCBVI algorithm (Azar et al., 2017), improving both the bonus terms and the regret analysis. Additionally, we compare our version of UCBVI with both its original version and the state-of-the-art MVP algorithm. Our empirical validation demonstrates that improving the multiplicative constants in the bounds has significant positive effects on the empirical performance of the algorithms.
Authors: Ali Baheri, Cecilia O. Alm
Abstract: We present a hierarchical neuro-symbolic control framework that tightly couples a classical symbolic planner with a transformer-based policy to address long-horizon decision-making under uncertainty. At the high level, the planner assembles an interpretable sequence of operators that guarantees logical coherence with task constraints, while at the low level each operator is rendered as a sub-goal token that conditions a decision transformer to generate fine-grained actions directly from raw observations. This bidirectional interface preserves the combinatorial efficiency and explainability of symbolic reasoning without sacrificing the adaptability of deep sequence models, and it permits a principled analysis that tracks how approximation errors from both planning and execution accumulate across the hierarchy. Empirical studies in stochastic grid-world domains demonstrate that the proposed method consistently surpasses purely symbolic, purely neural and existing hierarchical baselines in both success and efficiency, highlighting its robustness for sequential tasks.
Authors: Jaa-Yeon Lee, Byunghee Cha, Jeongsol Kim, Jong Chul Ye
Abstract: While recent advancements in generative modeling have significantly improved text-image alignment, some residual misalignment between text and image representations still remains. Some approaches address this issue by fine-tuning models in terms of preference optimization, etc., which require tailored datasets. Orthogonal to these methods, we revisit the challenge from the perspective of representation alignment-an approach that has gained popularity with the success of REPresentation Alignment (REPA). We first argue that conventional text-to-image (T2I) diffusion models, typically trained on paired image and text data (i.e., positive pairs) by minimizing score matching or flow matching losses, is suboptimal from the standpoint of representation alignment. Instead, a better alignment can be achieved through contrastive learning that leverages existing dataset as both positive and negative pairs. To enable efficient alignment with pretrained models, we propose SoftREPA- a lightweight contrastive fine-tuning strategy that leverages soft text tokens for representation alignment. This approach improves alignment with minimal computational overhead by adding fewer than 1M trainable parameters to the pretrained model. Our theoretical analysis demonstrates that our method explicitly increases the mutual information between text and image representations, leading to enhanced semantic consistency. Experimental results across text-to-image generation and text-guided image editing tasks validate the effectiveness of our approach in improving the semantic consistency of T2I generative models.
Authors: Yunli Wang, Zhen Zhang, Zhiqiang Wang, Zixuan Yang, Yu Li, Jian Yang, Shiyang Wen, Peng Jiang, Kun Gai
Abstract: Cascade Ranking is a prevalent architecture in large-scale top-k selection systems like recommendation and advertising platforms. Traditional training methods focus on single-stage optimization, neglecting interactions between stages. Recent advances have introduced interaction-aware training paradigms, but still struggle to 1) align training objectives with the goal of the entire cascade ranking (i.e., end-to-end recall of ground-truth items) and 2) learn effective collaboration patterns for different stages. To address these challenges, we propose LCRON, which introduces a novel surrogate loss function derived from the lower bound probability that ground truth items are selected by cascade ranking, ensuring alignment with the overall objective of the system. According to the properties of the derived bound, we further design an auxiliary loss for each stage to drive the reduction of this bound, leading to a more robust and effective top-k selection. LCRON enables end-to-end training of the entire cascade ranking system as a unified network. Experimental results demonstrate that LCRON achieves significant improvement over existing methods on public benchmarks and industrial applications, addressing key limitations in cascade ranking training and significantly enhancing system performance.
Authors: Haiyang Guo, Fanhu Zeng, Ziwei Xiang, Fei Zhu, Da-Han Wang, Xu-Yao Zhang, Cheng-Lin Liu
Abstract: Instruction tuning is widely used to improve a pre-trained Multimodal Large Language Model (MLLM) by training it on curated task-specific datasets, enabling better comprehension of human instructions. However, it is infeasible to collect all possible instruction datasets simultaneously in real-world scenarios. Thus, enabling MLLM with continual instruction tuning is essential for maintaining their adaptability. However, existing methods often trade off memory efficiency for performance gains, significantly compromising overall efficiency. In this paper, we propose a task-specific expansion and task-general fusion framework based on the variations in Centered Kernel Alignment (CKA) similarity across different model layers when trained on diverse datasets. Furthermore, we analyze the information leakage present in the existing benchmark and propose a new and more challenging benchmark to rationally evaluate the performance of different methods. Comprehensive experiments showcase a significant performance improvement of our method compared to existing state-of-the-art methods. Code and dataset are released at https://github.com/Ghy0501/HiDe-LLaVA.
Authors: Ibrahim Al Azher, Miftahul Jannat Mokarrama, Zhishuai Guo, Sagnik Ray Choudhury, Hamed Alhoori
Abstract: The future work section of a scientific article outlines potential research directions by identifying gaps and limitations of a current study. This section serves as a valuable resource for early-career researchers seeking unexplored areas and experienced researchers looking for new projects or collaborations. In this study, we generate future work suggestions from key sections of a scientific article alongside related papers and analyze how the trends have evolved. We experimented with various Large Language Models (LLMs) and integrated Retrieval-Augmented Generation (RAG) to enhance the generation process. We incorporate a LLM feedback mechanism to improve the quality of the generated content and propose an LLM-as-a-judge approach for evaluation. Our results demonstrated that the RAG-based approach with LLM feedback outperforms other methods evaluated through qualitative and quantitative metrics. Moreover, we conduct a human evaluation to assess the LLM as an extractor and judge. The code and dataset for this project are here, code: HuggingFace
Authors: Rochana Chaturvedi, Peyman Baghershahi, Sourav Medya, Barbara Di Eugenio
Abstract: Temporal information extraction from unstructured text is essential for contextualizing events and deriving actionable insights, particularly in the medical domain. We address the task of extracting clinical events and their temporal relations using the well-studied I2B2 2012 Temporal Relations Challenge corpus. This task is inherently challenging due to complex clinical language, long documents, and sparse annotations. We introduce GRAPHTREX, a novel method integrating span-based entity-relation extraction, clinical large pre-trained language models (LPLMs), and Heterogeneous Graph Transformers (HGT) to capture local and global dependencies. Our HGT component facilitates information propagation across the document through innovative global landmarks that bridge distant entities. Our method improves the state-of-the-art with 5.5% improvement in the tempeval $F_1$ score over the previous best and up to 8.9% improvement on long-range relations, which presents a formidable challenge. We further demonstrate generalizability by establishing a strong baseline on the E3C corpus. This work not only advances temporal information extraction but also lays the groundwork for improved diagnostic and prognostic models through enhanced temporal reasoning.
Authors: Arnav Arora, Srishti Yadav, Maria Antoniak, Serge Belongie, Isabelle Augenstein
Abstract: Automated frame analysis of political communication is a popular task in computational social science that is used to study how authors select aspects of a topic to frame its reception. So far, such studies have been narrow, in that they use a fixed set of pre-defined frames and focus only on the text, ignoring the visual contexts in which those texts appear. Especially for framing in the news, this leaves out valuable information about editorial choices, which include not just the written article but also accompanying photographs. To overcome such limitations, we present a method for conducting multi-modal, multi-label framing analysis at scale using large (vision-) language models. Grounding our work in framing theory, we extract latent meaning embedded in images used to convey a certain point and contrast that to the text by comparing the respective frames used. We also identify highly partisan framing of topics with issue-specific frame analysis found in prior qualitative work. We demonstrate a method for doing scalable integrative framing analysis of both text and image in news, providing a more complete picture for understanding media bias.
Authors: Dominik Schnaus, Nikita Araslanov, Daniel Cremers
Abstract: The platonic representation hypothesis suggests that vision and language embeddings become more homogeneous as model and dataset sizes increase. In particular, pairwise distances within each modality become more similar. This suggests that as foundation models mature, it may become possible to match vision and language embeddings in a fully unsupervised fashion, i.e. without parallel data. We present the first feasibility study, and investigate conformity of existing vision and language foundation models in the context of unsupervised, or "blind", matching. First, we formulate unsupervised matching as a quadratic assignment problem and introduce a novel heuristic that outperforms previous solvers. We also develop a technique to find optimal matching problems, for which a non-trivial match is very likely. Second, we conduct an extensive study deploying a range of vision and language models on four datasets. Our analysis reveals that for many problem instances, vision and language representations can be indeed matched without supervision. This finding opens up the exciting possibility of embedding semantic knowledge into other modalities virtually annotation-free. As a proof of concept, we showcase an unsupervised classifier, which achieves non-trivial classification accuracy without any image-text annotation.
Authors: Shuofei Qiao, Zhisong Qiu, Baochang Ren, Xiaobin Wang, Xiangyuan Ru, Ningyu Zhang, Xiang Chen, Yong Jiang, Pengjun Xie, Fei Huang, Huajun Chen
Abstract: Large Language Models (LLMs) have achieved considerable performance across various agentic planning tasks. However, traditional agent planning approaches adopt a "flood irrigation" methodology that indiscriminately injects gold trajectories, external feedback, and domain knowledge into agent models. This practice overlooks the fundamental human cognitive principle of situational self-awareness during decision-making-the ability to dynamically assess situational demands and strategically employ resources during decision-making. We propose agentic knowledgeable self-awareness to address this gap, a novel paradigm enabling LLM-based agents to autonomously regulate knowledge utilization. Specifically, we propose KnowSelf, a data-centric approach that applies agents with knowledgeable self-awareness like humans. Concretely, we devise a heuristic situation judgement criterion to mark special tokens on the agent's self-explored trajectories for collecting training data. Through a two-stage training process, the agent model can switch between different situations by generating specific special tokens, achieving optimal planning effects with minimal costs. Our experiments demonstrate that KnowSelf can outperform various strong baselines on different tasks and models with minimal use of external knowledge. Code is available at https://github.com/zjunlp/KnowSelf.
Authors: Zirui Liu, Jiatong Li, Yan Zhuang, Qi Liu, Shuanghong Shen, Jie Ouyang, Mingyue Cheng, Shijin Wang
Abstract: Arena-based evaluation is a fundamental yet significant evaluation paradigm for modern AI models, especially large language models (LLMs). Existing framework based on ELO rating system suffers from the inevitable instability problem due to ranking inconsistency and the lack of attention to the varying abilities of annotators. In this paper, we introduce a novel stable arena framework to address these issues by enhancing the ELO Rating System. Specifically, we replace the iterative update method with a Maximum Likelihood Estimation (MLE) approach, m-ELO, and provide theoretical proof of the consistency and stability of the MLE approach for model ranking. Additionally, we proposed the am-ELO, which modify the Elo Rating's probability function to incorporate annotator abilities, enabling the simultaneous estimation of model scores and annotator reliability. Experiments demonstrate that this method ensures stability, proving that this framework offers a more robust, accurate, and stable evaluation method for LLMs.
Authors: Soumik Dey, Hansi Wu, Binbin Li
Abstract: E-commerce sellers are recommended keyphrases based on their inventory on which they advertise to increase buyer engagement (clicks/sales). The relevance of advertiser keyphrases plays an important role in preventing the inundation of search systems with numerous irrelevant items that compete for attention in auctions, in addition to maintaining a healthy seller perception. In this work, we describe the shortcomings of training Advertiser keyphrase relevance filter models on click/sales/search relevance signals and the importance of aligning with human judgment, as sellers have the power to adopt or reject said keyphrase recommendations. In this study, we frame Advertiser keyphrase relevance as a complex interaction between 3 dynamical systems -- seller judgment, which influences seller adoption of our product, Advertising, which provides the keyphrases to bid on, and Search, who holds the auctions for the same keyphrases. This study discusses the practicalities of using human judgment via a case study at eBay Advertising and demonstrate that using LLM-as-a-judge en-masse as a scalable proxy for seller judgment to train our relevance models achieves a better harmony across the three systems -- provided that they are bound by a meticulous evaluation framework grounded in business metrics.
Authors: Yu Fan, Jingwei Ni, Jakob Merane, Etienne Salimbeni, Yang Tian, Yoan Hermstr\"uwer, Yinya Huang, Mubashara Akhtar, Florian Geering, Oliver Dreyer, Daniel Brunner, Markus Leippold, Mrinmaya Sachan, Alexander Stremitzer, Christoph Engel, Elliott Ash, Joel Niklaus
Abstract: Long-form legal reasoning remains a key challenge for large language models (LLMs) in spite of recent advances in test-time scaling. We introduce LEXam, a novel benchmark derived from 340 law exams spanning 116 law school courses across a range of subjects and degree levels. The dataset comprises 4,886 law exam questions in English and German, including 2,841 long-form, open-ended questions and 2,045 multiple-choice questions. Besides reference answers, the open questions are also accompanied by explicit guidance outlining the expected legal reasoning approach such as issue spotting, rule recall, or rule application. Our evaluation on both open-ended and multiple-choice questions present significant challenges for current LLMs; in particular, they notably struggle with open questions that require structured, multi-step legal reasoning. Moreover, our results underscore the effectiveness of the dataset in differentiating between models with varying capabilities. Adopting an LLM-as-a-Judge paradigm with rigorous human expert validation, we demonstrate how model-generated reasoning steps can be evaluated consistently and accurately. Our evaluation setup provides a scalable method to assess legal reasoning quality beyond simple accuracy metrics. Project page: https://lexam-benchmark.github.io/
Authors: Hao Dong, Ziyue Qiao, Zhiyuan Ning, Qi Hao, Yi Du, Pengyang Wang, Yuanchun Zhou
Abstract: Temporal Knowledge Graphs (TKGs), as an extension of static Knowledge Graphs (KGs), incorporate the temporal feature to express the transience of knowledge by describing when facts occur. TKG extrapolation aims to infer possible future facts based on known history, which has garnered significant attention in recent years. Some existing methods treat TKG as a sequence of independent subgraphs to model temporal evolution patterns, demonstrating impressive reasoning performance. However, they still have limitations: 1) In modeling subgraph semantic evolution, they usually neglect the internal structural interactions between subgraphs, which are actually crucial for encoding TKGs. 2) They overlook the potential smooth features that do not lead to semantic changes, which should be distinguished from the semantic evolution process. Therefore, we propose a novel Disentangled Multi-span Evolutionary Network (DiMNet) for TKG reasoning. Specifically, we design a multi-span evolution strategy that captures local neighbor features while perceiving historical neighbor semantic information, thus enabling internal interactions between subgraphs during the evolution process. To maximize the capture of semantic change patterns, we design a disentangle component that adaptively separates nodes' active and stable features, used to dynamically control the influence of historical semantics on future evolution. Extensive experiments conducted on four real-world TKG datasets show that DiMNet demonstrates substantial performance in TKG reasoning, and outperforms the state-of-the-art up to 22.7% in MRR.
Authors: Sangyong Lee, Subo Hwang, Dohoon Kim
Abstract: In this paper, we propose MADCluster, a novel model-agnostic anomaly detection framework utilizing self-supervised clustering. MADCluster is applicable to various deep learning architectures and addresses the 'hypersphere collapse' problem inherent in existing deep learning-based anomaly detection methods. The core idea is to cluster normal pattern data into a 'single cluster' while simultaneously learning the cluster center and mapping data close to this center. Also, to improve expressiveness and enable effective single clustering, we propose a new 'One-directed Adaptive loss'. The optimization of this loss is mathematically proven. MADCluster consists of three main components: Base Embedder capturing high-dimensional temporal dynamics, Cluster Distance Mapping, and Sequence-wise Clustering for continuous center updates. Its model-agnostic characteristics are achieved by applying various architectures to the Base Embedder. Experiments on four time series benchmark datasets demonstrate that applying MADCluster improves the overall performance of comparative models. In conclusion, the compatibility of MADCluster shows potential for enhancing model performance across various architectures.
Authors: Ahmed Heakl, Sarim Hashmi, Gustavo Bertolo Stahl, Seung Hun Eddie Han, Salman Khan, Abdulrahman Mahmoud
Abstract: We introduce CASS, the first large-scale dataset and model suite for cross-architecture GPU code transpilation, targeting both source-level (CUDA <--> HIP) and assembly-level (Nvidia SASS <--> AMD RDNA3) translation. The dataset comprises 70k verified code pairs across host and device, addressing a critical gap in low-level GPU code portability. Leveraging this resource, we train the CASS family of domain-specific language models, achieving 95% source translation accuracy and 37.5% assembly translation accuracy, substantially outperforming commercial baselines such as GPT-4o, Claude, and Hipify. Our generated code matches native performance in over 85% of test cases, preserving runtime and memory behavior. To support rigorous evaluation, we introduce CASS-Bench, a curated benchmark spanning 16 GPU domains with ground-truth execution. All data, models, and evaluation tools are released as open source to foster progress in GPU compiler tooling, binary compatibility, and LLM-guided hardware translation.
Authors: Licheng Pan, Yongqi Tong, Xin Zhang, Xiaolu Zhang, Jun Zhou, Zhixuan Chu
Abstract: Large language models (LLMs) have demonstrated remarkable capabilities across a wide range of tasks, yet they often refuse to answer legitimate queries-a phenomenon known as overrefusal. Overrefusal typically stems from over-conservative safety alignment, causing models to treat many reasonable prompts as potentially risky. To systematically understand this issue, we probe and leverage the models'safety decision boundaries to analyze and mitigate overrefusal. Our findings reveal that overrefusal is closely tied to misalignment at these boundary regions, where models struggle to distinguish subtle differences between benign and harmful content. Building on these insights, we present RASS, an automated framework for prompt generation and selection that strategically targets overrefusal prompts near the safety boundary. By harnessing steering vectors in the representation space, RASS efficiently identifies and curates boundary-aligned prompts, enabling more effective and targeted mitigation of overrefusal. This approach not only provides a more precise and interpretable view of model safety decisions but also seamlessly extends to multilingual scenarios.We have explored the safety decision boundaries of various LLMs and construct the MORBench evaluation set to facilitate robust assessment of model safety and helpfulness across multiple languages. Code and datasets will be released at https://anonymous.4open.science/r/RASS-80D3.
Authors: Pingchuan Ma, Ziang Yin, Qi Jing, Zhengqi Gao, Nicholas Gangi, Boyang Zhang, Tsung-Wei Huang, Zhaoran Huang, Duane S. Boning, Yu Yao, Jiaqi Gu
Abstract: DONNs leverage light propagation for efficient analog AI and signal processing. Advances in nanophotonic fabrication and metasurface-based wavefront engineering have opened new pathways to realize high-capacity DONNs across various spectral regimes. Training such DONN systems to determine the metasurface structures remains challenging. Heuristic methods are fast but oversimplify metasurfaces modulation, often resulting in physically unrealizable designs and significant performance degradation. Simulation-in-the-loop optimizes implementable metasurfaces via adjoint methods, but is computationally prohibitive and unscalable. To address these limitations, we propose SP2RINT, a spatially decoupled, progressive training framework that formulates DONN training as a PDE-constrained learning problem. Metasurface responses are first relaxed into freely trainable transfer matrices with a banded structure. We then progressively enforce physical constraints by alternating between transfer matrix training and adjoint-based inverse design, avoiding per-iteration PDE solves while ensuring final physical realizability. To further reduce runtime, we introduce a physics-inspired, spatially decoupled inverse design strategy based on the natural locality of field interactions. This approach partitions the metasurface into independently solvable patches, enabling scalable and parallel inverse design with system-level calibration. Evaluated across diverse DONN training tasks, SP2RINT achieves digital-comparable accuracy while being 1825 times faster than simulation-in-the-loop approaches. By bridging the gap between abstract DONN models and implementable photonic hardware, SP2RINT enables scalable, high-performance training of physically realizable meta-optical neural systems. Our code is available at https://github.com/ScopeX-ASU/SP2RINT
Authors: Hong Jiao, Dan Song, Won-Chan Lee
Abstract: Large language models (LLMs) have been widely explored for automated scoring in low-stakes assessment to facilitate learning and instruction. Empirical evidence related to which LLM produces the most reliable scores and induces least rater effects needs to be collected before the use of LLMs for automated scoring in practice. This study compared ten LLMs (ChatGPT 3.5, ChatGPT 4, ChatGPT 4o, OpenAI o1, Claude 3.5 Sonnet, Gemini 1.5, Gemini 1.5 Pro, Gemini 2.0, as well as DeepSeek V3, and DeepSeek R1) with human expert raters in scoring two types of writing tasks. The accuracy of the holistic and analytic scores from LLMs compared with human raters was evaluated in terms of Quadratic Weighted Kappa. Intra-rater consistency across prompts was compared in terms of Cronbach Alpha. Rater effects of LLMs were evaluated and compared with human raters using the Many-Facet Rasch model. The results in general supported the use of ChatGPT 4o, Gemini 1.5 Pro, and Claude 3.5 Sonnet with high scoring accuracy, better rater reliability, and less rater effects.
Authors: Ashirbad Mishra, Jinyu Zhao, Soumik Dey, Hansi Wu, Binbin Li, Kamesh Madduri
Abstract: In the domain of sponsored search advertising, the focus of Keyphrase recommendation has largely been on exact match types, which pose issues such as high management expenses, limited targeting scope, and evolving search query patterns. Alternatives like Broad match types can alleviate certain drawbacks of exact matches but present challenges like poor targeting accuracy and minimal supervisory signals owing to limited advertiser usage. This research defines the criteria for an ideal broad match, emphasizing on both efficiency and effectiveness, ensuring that a significant portion of matched queries are relevant. We propose BroadGen, an innovative framework that recommends efficient and effective broad match keyphrases by utilizing historical search query data. Additionally, we demonstrate that BroadGen, through token correspondence modeling, maintains better query stability over time. BroadGen's capabilities allow it to serve daily, millions of sellers at eBay with over 2.3 billion items.
Authors: Manuela Gonz\'alez-Gonz\'alez, Soufiane Belharbi, Muhammad Osama Zeeshan, Masoumeh Sharafi, Muhammad Haseeb Aslam, Marco Pedersoli, Alessandro Lameiras Koerich, Simon L Bacon, Eric Granger
Abstract: Recognizing complex emotions linked to ambivalence and hesitancy (A/H) can play a critical role in the personalization and effectiveness of digital behaviour change interventions. These subtle and conflicting emotions are manifested by a discord between multiple modalities, such as facial and vocal expressions, and body language. Although experts can be trained to identify A/H, integrating them into digital interventions is costly and less effective. Automatic learning systems provide a cost-effective alternative that can adapt to individual users, and operate seamlessly within real-time, and resource-limited environments. However, there are currently no datasets available for the design of ML models to recognize A/H. This paper introduces a first Behavioural Ambivalence/Hesitancy (BAH) dataset collected for subject-based multimodal recognition of A/H in videos. It contains videos from 224 participants captured across 9 provinces in Canada, with different age, and ethnicity. Through our web platform, we recruited participants to answer 7 questions, some of which were designed to elicit A/H while recording themselves via webcam with microphone. BAH amounts to 1,118 videos for a total duration of 8.26 hours with 1.5 hours of A/H. Our behavioural team annotated timestamp segments to indicate where A/H occurs, and provide frame- and video-level annotations with the A/H cues. Video transcripts and their timestamps are also included, along with cropped and aligned faces in each frame, and a variety of participants meta-data. We include results baselines for BAH at frame- and video-level recognition in multi-modal setups, in addition to zero-shot prediction, and for personalization using unsupervised domain adaptation. The limited performance of baseline models highlights the challenges of recognizing A/H in real-world videos. The data, code, and pretrained weights are available.
Authors: Jihoon Lee, Min Song
Abstract: Despite significant advancements in Large Vision-Language Models, Object Hallucination (OH) remains a persistent challenge. Building upon prior studies on contrastive decoding that address this issue without requiring additional model training, we introduce RVCD (Retrieval Visual Contrastive Decoding), an advanced method to suppress OH. RVCD leverages both negative and positive images at the logit level, explicitly referencing AI-generated images designed to represent a single concept. Our approach demonstrates substantial improvements over existing decoding-based methods.
Authors: Taehyo Kim, Qiran Jia, Mony J. de Leon, Hai Shu
Abstract: False discovery rate (FDR) control methods are essential for voxel-wise multiple testing in neuroimaging data analysis, where hundreds of thousands or even millions of tests are conducted to detect brain regions associated with disease-related changes. Classical FDR control methods (e.g., BH, q-value, and LocalFDR) assume independence among tests and often lead to high false non-discovery rates (FNR). Although various spatial FDR control methods have been developed to improve power, they still fall short of jointly addressing three major challenges in neuroimaging applications: capturing complex spatial dependencies, maintaining low variability in both false discovery proportion (FDP) and false non-discovery proportion (FNP) across replications, and achieving computational scalability for high-resolution data. To address these challenges, we propose fcHMRF-LIS, a powerful, stable, and scalable spatial FDR control method for voxel-wise multiple testing. It integrates the local index of significance (LIS)-based testing procedure with a novel fully connected hidden Markov random field (fcHMRF) designed to model complex spatial structures using a parsimonious parameterization. We develop an efficient expectation-maximization algorithm incorporating mean-field approximation, the Conditional Random Fields as Recurrent Neural Networks (CRF-RNN) technique, and permutohedral lattice filtering, reducing the time complexity from quadratic to linear in the number of tests. Extensive simulations demonstrate that fcHMRF-LIS achieves accurate FDR control, lower FNR, reduced variability in FDP and FNP, and a higher number of true positives compared to existing methods. Applied to an FDG-PET dataset from the Alzheimer's Disease Neuroimaging Initiative, fcHMRF-LIS identifies neurobiologically relevant brain regions and offers notable advantages in computational efficiency.
Authors: Jingping Nie, Dung T. Tran, Karan Thakkar, Vasudha Kowtha, Jon Huang, Carlos Avendano, Erdrin Azemi, Vikramjit Mitra
Abstract: Auscultation, particularly heart sound, is a non-invasive technique that provides essential vital sign information. Recently, self-supervised acoustic representation foundation models (FMs) have been proposed to offer insights into acoustics-based vital signs. However, there has been little exploration of the extent to which auscultation is encoded in these pre-trained FM representations. In this work, using a publicly available phonocardiogram (PCG) dataset and a heart rate (HR) estimation model, we conduct a layer-wise investigation of six acoustic representation FMs: HuBERT, wav2vec2, wavLM, Whisper, Contrastive Language-Audio Pretraining (CLAP), and an in-house CLAP model. Additionally, we implement the baseline method from Nie et al., 2024 (which relies on acoustic features) and show that overall, representation vectors from pre-trained foundation models (FMs) offer comparable performance to the baseline. Notably, HR estimation using the representations from the audio encoder of the in-house CLAP model outperforms the results obtained from the baseline, achieving a lower mean absolute error (MAE) across various train/validation/test splits despite the domain mismatch.
Authors: Shiqi Zhang, Tuomas Virtanen
Abstract: Bioacoustic sound event detection (BioSED) is crucial for biodiversity conservation but faces practical challenges during model development and training: limited amounts of annotated data, sparse events, species diversity, and class imbalance. To address these challenges efficiently with a limited labeling budget, we apply the mismatch-first farthest-traversal (MFFT), an active learning method integrating committee voting disagreement and diversity analysis. We also refine an existing BioSED dataset specifically for evaluating active learning algorithms. Experimental results demonstrate that MFFT achieves a mAP of 68% when cold-starting and 71% when warm-starting (which is close to the fully-supervised mAP of 75%) while using only 2.3% of the annotations. Notably, MFFT excels in cold-start scenarios and with rare species, which are critical for monitoring endangered species, demonstrating its practical value.
Authors: Jingyuan Huang, Xi Zhu, Minghao Guo, Yongfeng Zhang
Abstract: Web 2.0 social platforms are inherently centralized, with user data and algorithmic decisions controlled by the platform. However, users can only passively receive social predictions without being able to choose the underlying algorithm, which limits personalization. Fortunately, with the emergence of blockchain, users are allowed to choose algorithms that are tailored to their local situation, improving prediction results in a personalized way. In a blockchain environment, each user possesses its own model to perform the social prediction, capturing different perspectives on social interactions. In our work, we propose DeSocial, a decentralized social network learning framework deployed on an Ethereum (ETH) local development chain that integrates distributed data storage, node-level consensus, and user-driven model selection through Ganache. In the first stage, each user leverages DeSocial to evaluate multiple backbone models on their local subgraph. DeSocial coordinates the execution and returns model-wise prediction results, enabling the user to select the most suitable backbone for personalized social prediction. Then, DeSocial uniformly selects several validation nodes that possess the algorithm specified by each user, and aggregates the prediction results by majority voting, to prevent errors caused by any single model's misjudgment. Extensive experiments show that DeSocial has an evident improvement compared to the five classical centralized social network learning models, promoting user empowerment in blockchain-based decentralized social networks, showing the importance of multi-node validation and personalized algorithm selection based on blockchain. Our implementation is available at: https://github.com/agiresearch/DeSocial.
Authors: Yuxin Ren, Maxwell D Collins, Miao Hu, Huanrui Yang
Abstract: While transformers excel across vision and language pretraining tasks, their reliance on attention mechanisms poses challenges for inference efficiency, especially on edge and embedded accelerators with limited parallelism and memory bandwidth. Hinted by the observed redundancy of attention at inference time, we hypothesize that though the model learns complicated token dependency through pretraining, the inference-time sequence-to-sequence mapping in each attention layer is actually ''simple'' enough to be represented with a much cheaper function. In this work, we explore FAR, a Function-preserving Attention Replacement framework that replaces all attention blocks in pretrained transformers with learnable sequence-to-sequence modules, exemplified by an LSTM. FAR optimize a multi-head LSTM architecture with a block-wise distillation objective and a global structural pruning framework to achieve a family of efficient LSTM-based models from pretrained transformers. We validate FAR on the DeiT vision transformer family and demonstrate that it matches the accuracy of the original models on ImageNet and multiple downstream tasks with reduced parameters and latency. Further analysis shows that FAR preserves the semantic token relationships and the token-to-token correlation learned in the transformer's attention module.
Authors: Ahmed Heakl, Yahia Salaheldin Shaaban, Martin Takac, Salem Lahlou, Zangir Iklassov
Abstract: Robust routing under uncertainty is central to real-world logistics, yet most benchmarks assume static, idealized settings. We present SVRPBench, the first open benchmark to capture high-fidelity stochastic dynamics in vehicle routing at urban scale. Spanning more than 500 instances with up to 1000 customers, it simulates realistic delivery conditions: time-dependent congestion, log-normal delays, probabilistic accidents, and empirically grounded time windows for residential and commercial clients. Our pipeline generates diverse, constraint-rich scenarios, including multi-depot and multi-vehicle setups. Benchmarking reveals that state-of-the-art RL solvers like POMO and AM degrade by over 20% under distributional shift, while classical and metaheuristic methods remain robust. To enable reproducible research, we release the dataset and evaluation suite. SVRPBench challenges the community to design solvers that generalize beyond synthetic assumptions and adapt to real-world uncertainty.
Authors: Mengdan Zhu, Senhao Cheng, Guangji Bai, Yifei Zhang, Liang Zhao
Abstract: Text-to-image generation increasingly demands access to domain-specific, fine-grained, and rapidly evolving knowledge that pretrained models cannot fully capture. Existing Retrieval-Augmented Generation (RAG) methods attempt to address this by retrieving globally relevant images, but they fail when no single image contains all desired elements from a complex user query. We propose Cross-modal RAG, a novel framework that decomposes both queries and images into sub-dimensional components, enabling subquery-aware retrieval and generation. Our method introduces a hybrid retrieval strategy - combining a sub-dimensional sparse retriever with a dense retriever - to identify a Pareto-optimal set of images, each contributing complementary aspects of the query. During generation, a multimodal large language model is guided to selectively condition on relevant visual features aligned to specific subqueries, ensuring subquery-aware image synthesis. Extensive experiments on MS-COCO, Flickr30K, WikiArt, CUB, and ImageNet-LT demonstrate that Cross-modal RAG significantly outperforms existing baselines in both retrieval and generation quality, while maintaining high efficiency.
Authors: Tonghe Zhang, Chao Yu, Sichang Su, Yu Wang
Abstract: We propose ReinFlow, a simple yet effective online reinforcement learning (RL) framework that fine-tunes a family of flow matching policies for continuous robotic control. Derived from rigorous RL theory, ReinFlow injects learnable noise into a flow policy's deterministic path, converting the flow into a discrete-time Markov Process for exact and straightforward likelihood computation. This conversion facilitates exploration and ensures training stability, enabling ReinFlow to fine-tune diverse flow model variants, including Rectified Flow [35] and Shortcut Models [19], particularly at very few or even one denoising step. We benchmark ReinFlow in representative locomotion and manipulation tasks, including long-horizon planning with visual input and sparse reward. The episode reward of Rectified Flow policies obtained an average net growth of 135.36% after fine-tuning in challenging legged locomotion tasks while saving denoising steps and 82.63% of wall time compared to state-of-the-art diffusion RL fine-tuning method DPPO [43]. The success rate of the Shortcut Model policies in state and visual manipulation tasks achieved an average net increase of 40.34% after fine-tuning with ReinFlow at four or even one denoising step, whose performance is comparable to fine-tuned DDIM policies while saving computation time for an average of 23.20%. Project webpage: https://reinflow.github.io/
Authors: Yudi Zhang, Weilin Zhao, Xu Han, Tiejun Zhao, Wang Xu, Hailong Cao, Conghui Zhu
Abstract: Speculative decoding and quantization effectively accelerate memory-bound inference of large language models. Speculative decoding mitigates the memory bandwidth bottleneck by verifying multiple tokens within a single forward pass, which increases computational effort. Quantization achieves this optimization by compressing weights and activations into lower bit-widths and also reduces computations via low-bit matrix multiplications. To further leverage their strengths, we investigate the integration of these two techniques. Surprisingly, experiments applying the advanced speculative decoding method EAGLE-2 to various quantized models reveal that the memory benefits from 4-bit weight quantization are diminished by the computational load from speculative decoding. Specifically, verifying a tree-style draft incurs significantly more time overhead than a single-token forward pass on 4-bit weight quantized models. This finding led to our new speculative decoding design: a hierarchical framework that employs a small model as an intermediate stage to turn tree-style drafts into sequence drafts, leveraging the memory access benefits of the target quantized model. Experimental results show that our hierarchical approach achieves a 2.78$\times$ speedup across various tasks for the 4-bit weight Llama-3-70B model on an A100 GPU, outperforming EAGLE-2 by 1.31$\times$. Code available at https://github.com/AI9Stars/SpecMQuant.
Authors: Jiaxi Yang, Mengqi Zhang, Yiqiao Jin, Hao Chen, Qingsong Wen, Lu Lin, Yi He, Weijie Xu, James Evans, Jindong Wang
Abstract: Large Language Model-based Multi-Agent Systems (MASs) have emerged as a powerful paradigm for tackling complex tasks through collaborative intelligence. Nevertheless, the question of how agents should be structurally organized for optimal cooperation remains largely unexplored. In this position paper, we aim to gently redirect the focus of the MAS research community toward this critical dimension: develop topology-aware MASs for specific tasks. Specifically, the system consists of three core components - agents, communication links, and communication patterns - that collectively shape its coordination performance and efficiency. To this end, we introduce a systematic, three-stage framework: agent selection, structure profiling, and topology synthesis. Each stage would trigger new research opportunities in areas such as language models, reinforcement learning, graph learning, and generative modeling; together, they could unleash the full potential of MASs in complicated real-world applications. Then, we discuss the potential challenges and opportunities in the evaluation of multiple systems. We hope our perspective and framework can offer critical new insights in the era of agentic AI.
Authors: Mohamad Hakam Shams Eddin, Yikui Zhang, Stefan Kollet, Juergen Gall
Abstract: Recent deep learning approaches for river discharge forecasting have improved the accuracy and efficiency in flood forecasting, enabling more reliable early warning systems for risk management. Nevertheless, existing deep learning approaches in hydrology remain largely confined to local-scale applications and do not leverage the inherent spatial connections of bodies of water. Thus, there is a strong need for new deep learning methodologies that are capable of modeling spatio-temporal relations to improve river discharge and flood forecasting for scientific and operational applications. To address this, we present RiverMamba, a novel deep learning model that is pretrained with long-term reanalysis data and that can forecast global river discharge and floods on a $0.05^\circ$ grid up to 7 days lead time, which is of high relevance in early warning. To achieve this, RiverMamba leverages efficient Mamba blocks that enable the model to capture global-scale channel network routing and enhance its forecast capability for longer lead times. The forecast blocks integrate ECMWF HRES meteorological forecasts, while accounting for their inaccuracies through spatio-temporal modeling. Our analysis demonstrates that RiverMamba delivers reliable predictions of river discharge, including extreme floods across return periods and lead times, surpassing both operational AI- and physics-based models.
Authors: Luca Maria Del Bono, Federico Ricci-Tersenghi, Francesco Zamponi
Abstract: Recent years have seen a rise in the application of machine learning techniques to aid the simulation of hard-to-sample systems that cannot be studied using traditional methods. Despite the introduction of many different architectures and procedures, a wide theoretical understanding is still lacking, with the risk of suboptimal implementations. As a first step to address this gap, we provide here a complete analytic study of the widely-used Sequential Tempering procedure applied to a shallow MADE architecture for the Curie-Weiss model. The contribution of this work is twofold: firstly, we give a description of the optimal weights and of the training under Gradient Descent optimization. Secondly, we compare what happens in Sequential Tempering with and without the addition of local Metropolis Monte Carlo steps. We are thus able to give theoretical predictions on the best procedure to apply in this case. This work establishes a clear theoretical basis for the integration of machine learning techniques into Monte Carlo sampling and optimization.
Authors: Younggyo Seo, Carmelo Sferrazza, Haoran Geng, Michal Nauman, Zhao-Heng Yin, Pieter Abbeel
Abstract: Reinforcement learning (RL) has driven significant progress in robotics, but its complexity and long training times remain major bottlenecks. In this report, we introduce FastTD3, a simple, fast, and capable RL algorithm that significantly speeds up training for humanoid robots in popular suites such as HumanoidBench, IsaacLab, and MuJoCo Playground. Our recipe is remarkably simple: we train an off-policy TD3 agent with several modifications -- parallel simulation, large-batch updates, a distributional critic, and carefully tuned hyperparameters. FastTD3 solves a range of HumanoidBench tasks in under 3 hours on a single A100 GPU, while remaining stable during training. We also provide a lightweight and easy-to-use implementation of FastTD3 to accelerate RL research in robotics.
Authors: Guoxuan Chen, Lianghao Xia, Chao Huang
Abstract: Modern recommender systems powered by Graph Neural Networks (GNNs) excel at modeling complex user-item interactions, yet increasingly face scenarios requiring selective forgetting of training data. Beyond user requests to remove specific interactions due to privacy concerns or preference changes, regulatory frameworks mandate recommender systems' ability to eliminate the influence of certain user data from models. This recommendation unlearning challenge presents unique difficulties as removing connections within interaction graphs creates ripple effects throughout the model, potentially impacting recommendations for numerous users. Traditional approaches suffer from significant drawbacks: fragmentation methods damage graph structure and diminish performance, while influence function techniques make assumptions that may not hold in complex GNNs, particularly with self-supervised or random architectures. To address these limitations, we propose a novel model-agnostic pre-training paradigm UnlearnRec that prepares systems for efficient unlearning operations. Our Influence Encoder takes unlearning requests together with existing model parameters and directly produces updated parameters of unlearned model with little fine-tuning, avoiding complete retraining while preserving model performance characteristics. Extensive evaluation on public benchmarks demonstrates that our method delivers exceptional unlearning effectiveness while providing more than 10x speedup compared to retraining approaches. We release our method implementation at: https://github.com/HKUDS/UnlearnRec.