Authors: Yidong Chai (School of Management, Hefei University of Technology, Hefei, China, Key Laboratory of Process Optimization and Intelligent Decision-making, Ministry of Education, Hefei, China), Yang Liu (School of Management, Hefei University of Technology, Hefei, China, Key Laboratory of Process Optimization and Intelligent Decision-making, Ministry of Education, Hefei, China), Yonghang Zhou (School of Management, Hefei University of Technology, Hefei, China, Key Laboratory of Process Optimization and Intelligent Decision-making, Ministry of Education, Hefei, China), Jiaheng Xie (Department of Accounting and MIS, Lerner College of Business and Economics, University of Delaware, Newark, Delaware, U.S), Daniel Dajun Zeng (Institute of Automation, Chinese Academy of Sciences, Beijing, China)
Abstract: Large Language Models (LLMs) have demonstrated transformative potential in reshaping the world. As these models are pretrained on general corpora, they often require domain-specific fine-tuning to optimize performance in specialized business applications. Due to their massive scale, parameter-efficient fine-tuning (PEFT) methods are widely used to reduce training costs. Among them, hybrid PEFT methods that combine multiple PEFT techniques have achieved the best performance. However, existing hybrid PEFT methods face two main challenges when fine-tuning LLMs for specialized applications: (1) relying on point estimates, lacking the ability to quantify uncertainty for reliable decision-making, and (2) struggling to dynamically adapt to emerging data, lacking the ability to suit real-world situations. We propose Bayesian Hybrid Parameter-Efficient Fine-Tuning (BH-PEFT), a novel method that integrates Bayesian learning into hybrid PEFT. BH-PEFT combines Adapter, LoRA, and prefix-tuning to fine-tune feedforward and attention layers of the Transformer. By modeling learnable parameters as distributions, BH-PEFT enables uncertainty quantification. We further propose a Bayesian dynamic fine-tuning approach where the last posterior serves as the prior for the next round, enabling effective adaptation to new data. We evaluated BH-PEFT on business tasks such as sentiment analysis, news categorization, and commonsense reasoning. Results show that our method outperforms existing PEFT baselines, enables uncertainty quantification for more reliable decisions, and improves adaptability in dynamic scenarios. This work contributes to business analytics and data science by proposing a novel BH-PEFT method and dynamic fine-tuning approach that support uncertainty-aware and adaptive decision-making in real-world situations.
Authors: Samiksha BC
Abstract: This work introduces ZetA, a novel deep learning optimizer that extends Adam by incorporating dynamic scaling based on the Riemann zeta function. To the best of our knowledge, ZetA is the first optimizer to apply zeta-based gradient scaling within deep learning optimization. The method improves generalization and robustness through a hybrid update mechanism that integrates adaptive damping, cosine similarity-based momentum boosting, entropy-regularized loss, and Sharpness-Aware Minimization (SAM)-style perturbations. Empirical evaluations on SVHN, CIFAR10, CIFAR100, STL10, and noisy CIFAR10 consistently show test accuracy improvements over Adam. All experiments employ a lightweight fully connected network trained for five epochs under mixed-precision settings. The results demonstrate that ZetA is a computationally efficient and robust alternative to Adam, particularly effective in noisy or high-granularity classification tasks.
Authors: Yongfan Lai, Bo Liu, Xinyan Guan, Qinghao Zhao, Hongyan Li, Shenda Hong
Abstract: Personalized electrocardiogram (ECG) generation is to simulate a patient's ECG digital twins tailored to specific conditions. It has the potential to transform traditional healthcare into a more accurate individualized paradigm, while preserving the key benefits of conventional population-level ECG synthesis. However, this promising task presents two fundamental challenges: extracting individual features without ground truth and injecting various types of conditions without confusing generative model. In this paper, we present ECGTwin, a two-stage framework designed to address these challenges. In the first stage, an Individual Base Extractor trained via contrastive learning robustly captures personal features from a reference ECG. In the second stage, the extracted individual features, along with a target cardiac condition, are integrated into the diffusion-based generation process through our novel AdaX Condition Injector, which injects these signals via two dedicated and specialized pathways. Both qualitative and quantitative experiments have demonstrated that our model can not only generate ECG signals of high fidelity and diversity by offering a fine-grained generation controllability, but also preserving individual-specific features. Furthermore, ECGTwin shows the potential to enhance ECG auto-diagnosis in downstream application, confirming the possibility of precise personalized healthcare solutions.
Authors: Haitz S\'aez de Oc\'ariz Borde, Michael Bronstein
Abstract: We review the key mathematical concepts necessary for studying Geometric Deep Learning.
Authors: Md Imtiaz Habib
Abstract: In this research, I explore advanced deep learning methodologies to forecast the outcomes of the 2025 NCAA Division 1 Men's and Women's Basketball tournaments. Leveraging historical NCAA game data, I implement two sophisticated sequence-based models: Long Short-Term Memory (LSTM) and Transformer architectures. The predictive power of these models is augmented through comprehensive feature engineering, including team quality metrics derived from Generalized Linear Models (GLM), Elo ratings, seed differences, and aggregated box-score statistics. To evaluate the robustness and reliability of predictions, I train each model variant using both Binary Cross-Entropy (BCE) and Brier loss functions, providing insights into classification performance and probability calibration. My comparative analysis reveals that while the Transformer architecture optimized with BCE yields superior discriminative power (highest AUC of 0.8473), the LSTM model trained with Brier loss demonstrates superior probabilistic calibration (lowest Brier score of 0.1589). These findings underscore the importance of selecting appropriate model architectures and loss functions based on the specific requirements of forecasting tasks. The detailed analytical pipeline presented here serves as a reproducible framework for future predictive modeling tasks in sports analytics and beyond.
Authors: Tasnia Nobi Afee, Jack Hutchins, Md Mazharul Islam, Thomas Kampfe, Ahmedullah Aziz
Abstract: FeFETs hold strong potential for advancing memory and logic technologies, but their inherent randomness arising from both operational cycling and fabrication variability poses significant challenges for accurate and reliable modeling. Capturing this variability is critical, as it enables designers to predict behavior, optimize performance, and ensure reliability and robustness against variations in manufacturing and operating conditions. Existing deterministic and machine learning-based compact models often fail to capture the full extent of this variability or lack the mathematical smoothness required for stable circuit-level integration. In this work, we present an enhanced probabilistic modeling framework for FeFETs that addresses these limitations. Building upon a Mixture Density Network (MDN) foundation, our approach integrates C-infinity continuous activation functions for smooth, stable learning and a device-specific embedding layer to capture intrinsic physical variability across devices. Sampling from the learned embedding distribution enables the generation of synthetic device instances for variability-aware simulation. With an R2 of 0.92, the model demonstrates high accuracy in capturing the variability of FeFET current behavior. Altogether, this framework provides a scalable, data-driven solution for modeling the full stochastic behavior of FeFETs and offers a strong foundation for future compact model development and circuit simulation integration.
Authors: Zhixiang Lu, Yulong Li, Feilong Tang, Zhengyong Jiang, Chong Li, Mian Zhou, Tenglong Li, Jionglong Su
Abstract: Large-scale tuberculosis (TB) screening is limited by the high cost and operational complexity of traditional diagnostics, creating a need for artificial-intelligence solutions. We propose DeepGB-TB, a non-invasive system that instantly assigns TB risk scores using only cough audio and basic demographic data. The model couples a lightweight one-dimensional convolutional neural network for audio processing with a gradient-boosted decision tree for tabular features. Its principal innovation is a Cross-Modal Bidirectional Cross-Attention module (CM-BCA) that iteratively exchanges salient cues between modalities, emulating the way clinicians integrate symptoms and risk factors. To meet the clinical priority of minimizing missed cases, we design a Tuberculosis Risk-Balanced Loss (TRBL) that places stronger penalties on false-negative predictions, thereby reducing high-risk misclassifications. DeepGB-TB is evaluated on a diverse dataset of 1,105 patients collected across seven countries, achieving an AUROC of 0.903 and an F1-score of 0.851, representing a new state of the art. Its computational efficiency enables real-time, offline inference directly on common mobile devices, making it ideal for low-resource settings. Importantly, the system produces clinically validated explanations that promote trust and adoption by frontline health workers. By coupling AI innovation with public-health requirements for speed, affordability, and reliability, DeepGB-TB offers a tool for advancing global TB control.
Authors: Lu Gao, Ke Yu, Pan Lu
Abstract: Pavement deterioration modeling is important in providing information regarding the future state of the road network and in determining the needs of preventive maintenance or rehabilitation treatments. This research incorporated spatial dependence of road network into pavement deterioration modeling through a graph neural network (GNN). The key motivation of using a GNN for pavement performance modeling is the ability to easily and directly exploit the rich structural information in the network. This paper explored if considering spatial structure of the road network will improve the prediction performance of the deterioration models. The data used in this research comprises a large pavement condition data set with more than a half million observations taken from the Pavement Management Information System (PMIS) maintained by the Texas Department of Transportation. The promising comparison results indicates that pavement deterioration prediction models perform better when spatial relationship is considered.
Authors: Haoran Liu, Yihan Zhan, Mingzhe Liu, Yanhua Liu, Peng Li, Zhuo Zuo, Bingqi Liu, Runxi Liu
Abstract: This review presents a comprehensive survey and benchmark of pulse shape discrimination (PSD) algorithms for radiation detection, classifying nearly sixty methods into statistical (time-domain, frequency-domain, neural network-based) and prior-knowledge (machine learning, deep learning) paradigms. We implement and evaluate all algorithms on two standardized datasets: an unlabeled set from a 241Am-9Be source and a time-of-flight labeled set from a 238Pu-9Be source, using metrics including Figure of Merit (FOM), F1-score, ROC-AUC, and inter-method correlations. Our analysis reveals that deep learning models, particularly Multi-Layer Perceptrons (MLPs) and hybrid approaches combining statistical features with neural regression, often outperform traditional methods. We discuss architectural suitabilities, the limitations of FOM, alternative evaluation metrics, and performance across energy thresholds. Accompanying this work, we release an open-source toolbox in Python and MATLAB, along with the datasets, to promote reproducibility and advance PSD research.
Authors: Yi Zhao, Yajuan Peng, Cam-Tu Nguyen, Zuchao Li, Xiaoliang Wang, Hai Zhao, Xiaoming Fu
Abstract: KV cache eviction has emerged as an effective solution to alleviate resource constraints faced by LLMs in long-context scenarios. However, existing token-level eviction methods often overlook two critical aspects: (1) their irreversible eviction strategy fails to adapt to dynamic attention patterns during decoding (the saliency shift problem), and (2) they treat both marginally important tokens and truly unimportant tokens equally, despite the collective significance of marginal tokens to model performance (the marginal information over-compression problem). To address these issues, we design two compensation mechanisms based on the high similarity of attention matrices between LLMs of different scales. We propose SmallKV, a small model assisted compensation method for KV cache compression. SmallKV can maintain attention matching between different-scale LLMs to: 1) assist the larger model in perceiving globally important information of attention; and 2) use the smaller model's attention scores to approximate those of marginal tokens in the larger model. Extensive experiments on benchmarks including GSM8K, BBH, MT-Bench, and LongBench demonstrate the effectiveness of SmallKV. Moreover, efficiency evaluations show that SmallKV achieves 1.75 - 2.56 times higher throughput than baseline methods, highlighting its potential for efficient and performant LLM inference in resource constrained environments.
Authors: Haonan Yang, Jianchao Tang, Zhuo Li, Long Lan
Abstract: Time Series Forecasting (TSF) faces persistent challenges in modeling intricate temporal dependencies across different scales. Despite recent advances leveraging different decomposition operations and novel architectures based on CNN, MLP or Transformer, existing methods still struggle with static decomposition strategies, fragmented dependency modeling, and inflexible fusion mechanisms, limiting their ability to model intricate temporal dependencies. To explicitly solve the mentioned three problems respectively, we propose a novel Dynamic Multi-Scale Coordination Framework (DMSC) with Multi-Scale Patch Decomposition block (EMPD), Triad Interaction Block (TIB) and Adaptive Scale Routing MoE block (ASR-MoE). Specifically, EMPD is designed as a built-in component to dynamically segment sequences into hierarchical patches with exponentially scaled granularities, eliminating predefined scale constraints through input-adaptive patch adjustment. TIB then jointly models intra-patch, inter-patch, and cross-variable dependencies within each layer's decomposed representations. EMPD and TIB are jointly integrated into layers forming a multi-layer progressive cascade architecture, where coarse-grained representations from earlier layers adaptively guide fine-grained feature extraction in subsequent layers via gated pathways. And ASR-MoE dynamically fuses multi-scale predictions by leveraging specialized global and local experts with temporal-aware weighting. Comprehensive experiments on thirteen real-world benchmarks demonstrate that DMSC consistently maintains state-of-the-art (SOTA) performance and superior computational efficiency for TSF tasks. Code is available at https://github.com/1327679995/DMSC.
Authors: Dahun Kim, Anelia Angelova
Abstract: We propose Context-Adaptive Multi-Prompt Embedding, a novel approach to enrich semantic representations in vision-language contrastive learning. Unlike standard CLIP-style models that rely on a single text embedding, our method introduces multiple structured prompts, each containing a distinct adaptive token that captures diverse semantic aspects of the input text. We leverage a pretrained LLM as the text encoder within the CLIP framework, processing all prompts jointly in a single forward pass. The resulting prompt embeddings are combined into a unified text representation, enabling semantically richer alignment with visual features. To further promote semantic diversity and representation quality, we incorporate a diversity regularization loss and a negation-aware loss, encouraging specialization across prompts and improving contrastive discrimination. Our method achieves consistent improvements on both image-text and video-text retrieval benchmarks.
Authors: Oc\'eane Doremus, Ariel Guerra-Adames, Marta Avalos-Fernandez, Vianney Jouhet, C\'edric Gil-Jardin\'e, Emmanuel Lagarde
Abstract: Faced with the challenges of patient confidentiality and scientific reproducibility, research on machine learning for health is turning towards the conception of synthetic medical databases. This article presents a brief overview of state-of-the-art machine learning methods for generating synthetic tabular and textual data, focusing their application to the automatic classification of trauma mechanisms, followed by our proposed methodology for generating high-quality, synthetic medical records combining tabular and unstructured text data.
Authors: Katherine Avery, Chinmay Pendse, David Jensen
Abstract: Distributionally robust evaluation estimates the worst-case expected return over an uncertainty set of possible covariate and reward distributions, and distributionally robust learning finds a policy that maximizes that worst-case return across that uncertainty set. Unfortunately, current methods for distributionally robust evaluation and learning create overly conservative evaluations and policies. In this work, we propose a practical bandit evaluation and learning algorithm that tailors the uncertainty set to specific problems using mathematical programs constrained by structural equation models. Further, we show how conditional independence testing can be used to detect shifted variables for modeling. We find that the structural equation model (SEM) approach gives more accurate evaluations and learns lower-variance policies than traditional approaches, particularly for large shifts. Further, the SEM approach learns an optimal policy, assuming the model is sufficiently well-specified.
Authors: Lei Pang, Ruinan Jin
Abstract: Group Relative Policy Optimization (GRPO), recently proposed by DeepSeek, is a critic-free reinforcement learning algorithm for fine tuning large language models. It replaces the value function in Proximal Policy Optimization (PPO) with group normalized rewards, while retaining PPO style token level importance sampling based on an old policy. We show that GRPO update rule in fact estimates the policy gradient at the old policy rather than the current one. However, since the old policy is refreshed every few steps, the discrepancy between the two remains small limiting the impact of this bias in practice. We validate this through an ablation study in which importance sampling is entirely removed, and updates are instead performed using the gradient estimated at a fixed old policy across multiple optimization steps. Remarkably, this simplification results in performance comparable to standard GRPO. Motivated by these findings, we propose a new algorithm: Trajectory level Importance Corrected GRPO (TIC GRPO). TIC GRPO replaces token level importance ratios with a single trajectory level probability ratio, yielding an unbiased estimate of the current policy gradient while preserving the critic free structure. Furthermore, we present the first theoretical convergence analysis for GRPO style methods, covering both the original GRPO and our proposed variant.
Authors: Hanqi Feng, Peng Qiu, Mengchun Zhang, Yiran Tao, You Fan, Jingtao Xu, Barnabas Poczos
Abstract: Recent advances in diffusion models have shown remarkable potential for antibody design, yet existing approaches apply uniform generation strategies that cannot adapt to each antigen's unique requirements. Inspired by B cell affinity maturation, where antibodies evolve through multi-objective optimization balancing affinity, stability, and self-avoidance, we propose the first biologically-motivated framework that leverages physics-based domain knowledge within an online meta-learning system. Our method employs multiple specialized experts (van der Waals, molecular recognition, energy balance, and interface geometry) whose parameters evolve during generation based on iterative feedback, mimicking natural antibody refinement cycles. Instead of fixed protocols, this adaptive guidance discovers personalized optimization strategies for each target. Our experiments demonstrate that this approach: (1) discovers optimal SE(3)-equivariant guidance strategies for different antigen classes without pre-training, preserving molecular symmetries throughout optimization; (2) significantly enhances hotspot coverage and interface quality through target-specific adaptation, achieving balanced multi-objective optimization characteristic of therapeutic antibodies; (3) establishes a paradigm for iterative refinement where each antibody-antigen system learns its unique optimization profile through online evaluation; (4) generalizes effectively across diverse design challenges, from small epitopes to large protein interfaces, enabling precision-focused campaigns for individual targets.
Authors: Kennedy Edemacu, Vinay M. Shashidhar, Micheal Tuape, Dan Abudu, Beakcheol Jang, Jong Wook Kim
Abstract: Retrieval-Augmented Generation (RAG) has emerged as a powerful approach to boost the capabilities of large language models (LLMs) by incorporating external, up-to-date knowledge sources. However, this introduces a potential vulnerability to knowledge poisoning attacks, where attackers can compromise the knowledge source to mislead the generation model. One such attack is the PoisonedRAG in which the injected adversarial texts steer the model to generate an attacker-chosen response to a target question. In this work, we propose novel defense methods, FilterRAG and ML-FilterRAG, to mitigate the PoisonedRAG attack. First, we propose a new property to uncover distinct properties to differentiate between adversarial and clean texts in the knowledge data source. Next, we employ this property to filter out adversarial texts from clean ones in the design of our proposed approaches. Evaluation of these methods using benchmark datasets demonstrate their effectiveness, with performances close to those of the original RAG systems.
Authors: Chaoyang Gao, Xiang Chen, Jiyu Wang, Jibin Wang, Guang Yang
Abstract: The increasing complexity of software systems has led to a surge in cybersecurity vulnerabilities, necessitating efficient and scalable solutions for vulnerability assessment. However, the deployment of large pre-trained models in real-world scenarios is hindered by their substantial computational and storage demands. To address this challenge, we propose a novel resource-efficient framework that integrates knowledge distillation and particle swarm optimization to enable automated vulnerability assessment. Our framework employs a two-stage approach: First, particle swarm optimization is utilized to optimize the architecture of a compact student model, balancing computational efficiency and model capacity. Second, knowledge distillation is applied to transfer critical vulnerability assessment knowledge from a large teacher model to the optimized student model. This process significantly reduces the model size while maintaining high performance. Experimental results on an enhanced MegaVul dataset, comprising 12,071 CVSS (Common Vulnerability Scoring System) v3 annotated vulnerabilities, demonstrate the effectiveness of our approach. Our approach achieves a 99.4% reduction in model size while retaining 89.3% of the original model's accuracy. Furthermore, it outperforms state-of-the-art baselines by 1.7% in accuracy with 60% fewer parameters. The framework also reduces training time by 72.1% and architecture search time by 34.88% compared to traditional genetic algorithms.
Authors: Enrique Luna Villag\'omez, Vladimir Mahalec
Abstract: Kolmogorov-Arnold Networks (KANs) have recently emerged as a flexible and parameter-efficient alternative to conventional neural networks. Unlike standard architectures that use fixed node-based activations, KANs place learnable functions on edges, parameterized by different function families. While they have shown promise in supervised settings, their utility in unsupervised fault detection remains largely unexplored. This study presents a comparative evaluation of KAN-based autoencoders (KAN-AEs) for unsupervised fault detection in chemical processes. We investigate four KAN-AE variants, each based on a different KAN implementation (EfficientKAN, FastKAN, FourierKAN, and WavKAN), and benchmark them against an Orthogonal Autoencoder (OAE) on the Tennessee Eastman Process. Models are trained on normal operating data across 13 training set sizes and evaluated on 21 fault types, using Fault Detection Rate (FDR) as the performance metric. WavKAN-AE achieves the highest overall FDR ($\geq$92\%) using just 4,000 training samples and remains the top performer, even as other variants are trained on larger datasets. EfficientKAN-AE reaches $\geq$90\% FDR with only 500 samples, demonstrating robustness in low-data settings. FastKAN-AE becomes competitive at larger scales ($\geq$50,000 samples), while FourierKAN-AE consistently underperforms. The OAE baseline improves gradually but requires substantially more data to match top KAN-AE performance. These results highlight the ability of KAN-AEs to combine data efficiency with strong fault detection performance. Their use of structured basis functions suggests potential for improved model transparency, making them promising candidates for deployment in data-constrained industrial settings.
Authors: Roman Gutierrez, Tony Kai Tang, Isabel Gutierrez
Abstract: Robust regression techniques rely on least-squares optimization, which works well for Gaussian noise but fails in the presence of asymmetric structured noise. We propose a hybrid neural-symbolic architecture where a transformer encoder processes numerical sequences, a compression NN predicts symbolic parameters, and a fixed symbolic equation reconstructs the original sequence. Using synthetic data, the training objective is to recover the original sequence after adding asymmetric structured noise, effectively learning a symbolic fit guided by neural parameter estimation. Our model achieves a median regression MSE of 6e-6 to 3.5e-5 on synthetic wearable data, which is a 10-300 times improvement when compared with ordinary least squares fit and robust regression techniques such as Huber loss or SoftL1.
Authors: Shifeng Xie, Vasilii Feofanov, Marius Alonso, Ambroise Odonnat, Jianfeng Zhang, Themis Palpanas, Ievgen Redko
Abstract: Time series foundation models (TSFMs) have recently gained significant attention due to their strong zero-shot capabilities and widespread real-world applications. Such models typically require a computationally costly pretraining on large-scale, carefully curated collections of real-world sequences. To allow for a sample-efficient pretraining of TSFMs, we propose CauKer, a novel algorithm designed to generate diverse, causally coherent synthetic time series with realistic trends, seasonality, and nonlinear interactions. CauKer combines Gaussian Process (GP) kernel composition with Structural Causal Models (SCM) to produce data for sample-efficient pretraining of state-of-the-art classification TSFMs having different architectures and following different pretraining approaches. Additionally, our experiments reveal that CauKer-generated datasets exhibit clear scaling laws for both dataset size (10K to 10M samples) and model capacity (1M to 783M parameters), unlike real-world datasets, which display irregular scaling behavior.
Authors: Alex Massucco, Davide Murari, Carola-Bibiane Sch\"onlieb
Abstract: Very deep neural networks achieve state-of-the-art performance by extracting rich, hierarchical features. Yet, training them via backpropagation is often hindered by vanishing or exploding gradients. Existing remedies, such as orthogonal or variance-preserving initialisation and residual architectures, allow for a more stable gradient propagation and the training of deeper models. In this work, we introduce a unified mathematical framework that describes a broad class of nonlinear feedforward and residual networks, whose input-to-output Jacobian matrices are exactly orthogonal almost everywhere. Such a constraint forces the resulting networks to achieve perfect dynamical isometry and train efficiently despite being very deep. Our formulation not only recovers standard architectures as particular cases but also yields new designs that match the trainability of residual networks without relying on conventional skip connections. We provide experimental evidence that perfect Jacobian orthogonality at initialisation is sufficient to stabilise training and achieve competitive performance. We compare this strategy to networks regularised to maintain the Jacobian orthogonality and obtain comparable results. We further extend our analysis to a class of networks well-approximated by those with orthogonal Jacobians and introduce networks with Jacobians representing partial isometries. These generalized models are then showed to maintain the favourable trainability properties.
Authors: Jialin Zheng, Haoyu Wang, Yangbin Zeng, Di Mou, Xin Zhang, Hong Li, Sergio Vazquez, Leopoldo G. Franquelo
Abstract: Edge Digital Twins (EDTs) are crucial for monitoring and control of Power Electronics Systems (PES). However, existing modeling approaches struggle to consistently capture continuously evolving hybrid dynamics that are inherent in PES, degrading Sim-to-Real generalization on resource-constrained edge devices. To address these challenges, this paper proposes a Physics-Embedded Neural ODEs (PENODE) that (i) embeds the hybrid operating mechanism as an event automaton to explicitly govern discrete switching and (ii) injects known governing ODE components directly into the neural parameterization of unmodeled dynamics. This unified design yields a differentiable end-to-end trainable architecture that preserves physical interpretability while reducing redundancy, and it supports a cloud-to-edge toolchain for efficient FPGA deployment. Experimental results demonstrate that PENODE achieves significantly higher accuracy in benchmarks in white-box, gray-box, and black-box scenarios, with a 75% reduction in neuron count, validating that the proposed PENODE maintains physical interpretability, efficient edge deployment, and real-time control enhancement.
Authors: Aakash Gore, Prasanna Chaporkar
Abstract: We study a stochastic multi-armed bandit setting where arms are partitioned into known clusters, such that the mean rewards of arms within a cluster differ by at most a known threshold. While the clustering structure is known a priori, the arm means are unknown. This framework models scenarios where outcomes depend on multiple factors -- some with significant and others with minor influence -- such as online advertising, clinical trials, and wireless communication. We derive asymptotic lower bounds on the regret that improve upon the classical bound of Lai & Robbins (1985). We then propose Clus-UCB, an efficient algorithm that closely matches this lower bound asymptotically. Clus-UCB is designed to exploit the clustering structure and introduces a new index to evaluate an arm, which depends on other arms within the cluster. In this way, arms share information among each other. We present simulation results of our algorithm and compare its performance against KL-UCB and other well-known algorithms for bandits with dependent arms. Finally, we address some limitations of this work and conclude by mentioning possible future research.
Authors: Zhong Zhang, Francesco Topputo
Abstract: In trajectory design, fuel consumption and trajectory reachability are two key performance indicators for low-thrust missions. This paper proposes general-purpose pretrained neural networks to predict these metrics. The contributions of this paper are as follows: Firstly, based on the confirmation of the Scaling Law applicable to low-thrust trajectory approximation, the largest dataset is constructed using the proposed homotopy ray method, which aligns with mission-design-oriented data requirements. Secondly, the data are transformed into a self-similar space, enabling the neural network to adapt to arbitrary semi-major axes, inclinations, and central bodies. This extends the applicability beyond existing studies and can generalize across diverse mission scenarios without retraining. Thirdly, to the best of our knowledge, this work presents the current most general and accurate low-thrust trajectory approximator, with implementations available in C++, Python, and MATLAB. The resulting neural network achieves a relative error of 0.78% in predicting velocity increments and 0.63% in minimum transfer time estimation. The models have also been validated on a third-party dataset, multi-flyby mission design problem, and mission analysis scenario, demonstrating their generalization capability, predictive accuracy, and computational efficiency.
Authors: Biyi Fang, Jean Utke, Truong Vo, Diego Klabjan
Abstract: Transformer architectures dominate modern NLP but often demand heavy computational resources and intricate hyperparameter tuning. To mitigate these challenges, we propose a novel framework, BoostTransformer, that augments transformers with boosting principles through subgrid token selection and importance-weighted sampling. Our method incorporates a least square boosting objective directly into the transformer pipeline, enabling more efficient training and improved performance. Across multiple fine-grained text classification benchmarks, BoostTransformer demonstrates both faster convergence and higher accuracy, surpassing standard transformers while minimizing architectural search overhead.
Authors: Arthur Cho
Abstract: Generative Machine Learning models have become central to modern systems, powering applications in creative writing, summarization, multi-hop reasoning, and context-aware dialogue. These models underpin large-scale AI assistants, workflow automation, and autonomous decision-making. In such domains, acceptable response is rarely absolute or static, but plural and highly context-dependent. Yet standard evaluation regimes still rely on static, benchmark-style tests, incentivizing optimization toward leaderboard scores rather than alignment with dynamic user needs or evolving realities. GrandJury introduces a formal evaluation protocol combining time-decayed aggregation, complete traceability, with the support of dynamic, transparent task rubric attribution, and multi-rater human judgment. Together, these elements enable pluralistic, accountable evaluation that captures evolving consensus and surfaces disagreement. We provide an open-source implementation (grandjury PyPI package) and a public collection of Large Language Model (LLM) inference outputs to illustrate the need and method. GrandJury provides a new paradigm for AI practitioners when evaluating machine learning outputs without absolute ground truth.
Authors: Minghao Yan, Zhuang Wang, Zhen Jia, Shivaram Venkataraman, Yida Wang
Abstract: Low-rank Adaptation (LoRA) has gained popularity as a fine-tuning approach for Large Language Models (LLMs) due to its low resource requirements and good performance. While a plethora of work has investigated improving LoRA serving efficiency by serving multiple LoRAs concurrently, existing methods assume that a wide range of LoRA adapters are available for serving. In our work, we conduct extensive empirical studies to identify that current training paradigms do not utilize hardware resources efficiently and require high overhead to obtain a performant LoRA. Leveraging these insights, we propose PLoRA, which automatically orchestrates concurrent LoRA fine-tuning jobs under given hardware and model constraints and develops performant kernels to improve training efficiency. Our experimental studies show that PLoRA reduces the makespan of LoRA fine-tuning over a given hyperparameter search space by up to 7.52x and improves training throughput by up to 12.8x across a range of state-of-the-art LLMs.
Authors: Zain Ulabedeen Farhat, Debamita Ghosh, George K. Atia, Yue Wang
Abstract: Well-trained multi-agent systems can fail when deployed in real-world environments due to model mismatches between the training and deployment environments, caused by environment uncertainties including noise or adversarial attacks. Distributionally Robust Markov Games (DRMGs) enhance system resilience by optimizing for worst-case performance over a defined set of environmental uncertainties. However, current methods are limited by their dependence on simulators or large offline datasets, which are often unavailable. This paper pioneers the study of online learning in DRMGs, where agents learn directly from environmental interactions without prior data. We introduce the {\it Robust Optimistic Nash Value Iteration (RONAVI)} algorithm and provide the first provable guarantees for this setting. Our theoretical analysis demonstrates that the algorithm achieves low regret and efficiently finds the optimal robust policy for uncertainty sets measured by Total Variation divergence and Kullback-Leibler divergence. These results establish a new, practical path toward developing truly robust multi-agent systems.
Authors: Jonathan Patsenker, Henry Li, Myeongseob Ko, Ruoxi Jia, Yuval Kluger
Abstract: Diffusion models have been firmly established as principled zero-shot solvers for linear and nonlinear inverse problems, owing to their powerful image prior and iterative sampling algorithm. These approaches often rely on Tweedie's formula, which relates the diffusion variate $\mathbf{x}_t$ to the posterior mean $\mathbb{E} [\mathbf{x}_0 | \mathbf{x}_t]$, in order to guide the diffusion trajectory with an estimate of the final denoised sample $\mathbf{x}_0$. However, this does not consider information from the measurement $\mathbf{y}$, which must then be integrated downstream. In this work, we propose to estimate the conditional posterior mean $\mathbb{E} [\mathbf{x}_0 | \mathbf{x}_t, \mathbf{y}]$, which can be formulated as the solution to a lightweight, single-parameter maximum likelihood estimation problem. The resulting prediction can be integrated into any standard sampler, resulting in a fast and memory-efficient inverse solver. Our optimizer is amenable to a noise-aware likelihood-based stopping criteria that is robust to measurement noise in $\mathbf{y}$. We demonstrate comparable or improved performance against a wide selection of contemporary inverse solvers across multiple datasets and tasks.
Authors: Ninh Pham, Yingtao Zheng, Hugo Phibbs
Abstract: We propose a novel perspective on varied-density clustering for high-dimensional data by framing it as a label propagation process in neighborhood graphs that adapt to local density variations. Our method formally connects density-based clustering with graph connectivity, enabling the use of efficient graph propagation techniques developed in network science. To ensure scalability, we introduce a density-aware neighborhood propagation algorithm and leverage advanced random projection methods to construct approximate neighborhood graphs. Our approach significantly reduces computational cost while preserving clustering quality. Empirically, it scales to datasets with millions of points in minutes and achieves competitive accuracy compared to existing baselines.
Authors: Jiahui Bai, Hai Dong, A. K. Qin
Abstract: Decentralized Federated Learning (DFL) struggles with the slow adaptation of late-joining delayed clients and high communication costs in asynchronous environments. These limitations significantly hinder overall performance. To address this, we propose DFedCAD, a novel framework for rapid adaptation via Centroid-Aligned Distillation. DFedCAD first employs Weighted Cluster Pruning (WCP) to compress models into representative centroids, drastically reducing communication overhead. It then enables delayed clients to intelligently weigh and align with peer knowledge using a novel structural distance metric and a differentiable k-means distillation module, facilitating efficient end-to-end knowledge transfer. Extensive experiments on CIFAR-10, CIFAR-100, and Tiny-ImageNet show that DFedCAD consistently achieves state-of-the-art performance, attaining the highest accuracy across all evaluated settings while reducing communication overhead by over 86%. Our framework provides a scalable and practical solution for efficient decentralized learning in dynamic, real-world scenarios.
Authors: Haidong Kang, Lianbo Ma, Guo Yu, Shangce Gao
Abstract: Mixed precision quantization (MPQ) is an effective quantization approach to achieve accuracy-complexity trade-off of neural network, through assigning different bit-widths to network activations and weights in each layer. The typical way of existing MPQ methods is to optimize quantization policies (i.e., bit-width allocation) in a gradient descent manner, termed as Differentiable (DMPQ). At the end of the search, the bit-width associated to the quantization parameters which has the largest value will be selected to form the final mixed precision quantization policy, with the implicit assumption that the values of quantization parameters reflect the operation contribution to the accuracy improvement. While much has been discussed about the MPQ improvement, the bit-width selection process has received little attention. We study this problem and argue that the magnitude of quantization parameters does not necessarily reflect the actual contribution of the bit-width to the task performance. Then, we propose a Shapley-based MPQ (SMPQ) method, which measures the bit-width operation direct contribution on the MPQ task. To reduce computation cost, a Monte Carlo sampling-based approximation strategy is proposed for Shapley computation. Extensive experiments on mainstream benchmarks demonstrate that our SMPQ consistently achieves state-of-the-art performance than gradient-based competitors.
Authors: Ruixing Zhang, Bo Wang, Tongyu Zhu, Leilei Sun, Weifeng Lv
Abstract: Urban profiling aims to predict urban profiles in unknown regions and plays a critical role in economic and social censuses. Existing approaches typically follow a two-stage paradigm: first, learning representations of urban areas; second, performing downstream prediction via linear probing, which originates from the BERT era. Inspired by the development of GPT style models, recent studies have shown that novel self-supervised pretraining schemes can endow models with direct applicability to downstream tasks, thereby eliminating the need for task-specific fine-tuning. This is largely because GPT unifies the form of pretraining and inference through next-token prediction. However, urban data exhibit structural characteristics that differ fundamentally from language, making it challenging to design a one-stage model that unifies both pretraining and inference. In this work, we propose Urban In-Context Learning, a framework that unifies pretraining and inference via a masked autoencoding process over urban regions. To capture the distribution of urban profiles, we introduce the Urban Masked Diffusion Transformer, which enables each region' s prediction to be represented as a distribution rather than a deterministic value. Furthermore, to stabilize diffusion training, we propose the Urban Representation Alignment Mechanism, which regularizes the model's intermediate features by aligning them with those from classical urban profiling methods. Extensive experiments on three indicators across two cities demonstrate that our one-stage method consistently outperforms state-of-the-art two-stage approaches. Ablation studies and case studies further validate the effectiveness of each proposed module, particularly the use of diffusion modeling.
Authors: Tatwadarshi P Nagarhalli, Sanket Patil, Vishal Pande, Uday Aswalekar, Prafulla Patil
Abstract: Alzheimers Disease (AD) is a progressive neurodegenerative disorder that poses significant challenges in its early diagnosis, often leading to delayed treatment and poorer outcomes for patients. Traditional diagnostic methods, typically reliant on single data modalities, fall short of capturing the multifaceted nature of the disease. In this paper, we propose a novel multimodal framework for the early detection of AD that integrates data from three primary sources: MRI imaging, cognitive assessments, and biomarkers. This framework employs Convolutional Neural Networks (CNN) for analyzing MRI images and Long Short-Term Memory (LSTM) networks for processing cognitive and biomarker data. The system enhances diagnostic accuracy and reliability by aggregating results from these distinct modalities using advanced techniques like weighted averaging, even in incomplete data. The multimodal approach not only improves the robustness of the detection process but also enables the identification of AD at its earliest stages, offering a significant advantage over conventional methods. The integration of biomarkers and cognitive tests is particularly crucial, as these can detect Alzheimer's long before the onset of clinical symptoms, thereby facilitating earlier intervention and potentially altering the course of the disease. This research demonstrates that the proposed framework has the potential to revolutionize the early detection of AD, paving the way for more timely and effective treatments
Authors: Dingwei Zhu, Shihan Dou, Zhiheng Xi, Senjie Jin, Guoqiang Zhang, Jiazheng Zhang, Junjie Ye, Mingxu Chai, Enyu Zhou, Ming Zhang, Caishuang Huang, Yunke Zhang, Yuran Wang, Tao Gui
Abstract: Reinforcement Learning from Human Feedback (RLHF) often suffers from noisy or imperfect reward supervision in real-world settings, which undermines policy stability and generalization. Such noise may cause models to lose attention on key words during advantage estimation. While prior work focuses on reward denoising or filtering poor data, it often overlooks the critical role of the value model in policy optimization. In this work, we show that a strong value model is essential for mitigating noise by absorbing unstable signals and enabling more reliable advantage estimation. We propose VRPO, a value-centric framework for robust PPO training under noisy supervision. VRPO combines two core designs: (1) an auxiliary loss guided by entropy and perplexity from a frozen language model, and (2) a variational information bottleneck. These mechanisms enhance the value model's ability to filter out noise and capture key words from the context during advantage estimation, transforming it from a passive predictor into an active regulator of noise. Experiments on math reasoning, science QA, and multi-turn dialogue, under both rule-based and model-based noisy rewards, show that VRPO consistently outperforms PPO and GRPO baselines. Our findings underscore the often-overlooked importance of the value model in RLHF and offer a principled and practical approach to robust policy optimization in noisy real-world environments.
Authors: Sukruta Prakash Midigeshi, Tanmay Goyal, Gaurav Sinha
Abstract: Multinomial Logistic Bandits have recently attracted much attention due to their ability to model problems with multiple outcomes. In this setting, each decision is associated with many possible outcomes, modeled using a multinomial logit function. Several recent works on multinomial logistic bandits have simultaneously achieved optimal regret and computational efficiency. However, motivated by real-world challenges and practicality, there is a need to develop algorithms with limited adaptivity, wherein we are allowed only $M$ policy updates. To address these challenges, we present two algorithms, B-MNL-CB and RS-MNL, that operate in the batched and rarely-switching paradigms, respectively. The batched setting involves choosing the $M$ policy update rounds at the start of the algorithm, while the rarely-switching setting can choose these $M$ policy update rounds in an adaptive fashion. Our first algorithm, B-MNL-CB extends the notion of distributional optimal designs to the multinomial setting and achieves $\tilde{O}(\sqrt{T})$ regret assuming the contexts are generated stochastically when presented with $\Omega(\log \log T)$ update rounds. Our second algorithm, RS-MNL works with adversarially generated contexts and can achieve $\tilde{O}(\sqrt{T})$ regret with $\tilde{O}(\log T)$ policy updates. Further, we conducted experiments that demonstrate that our algorithms (with a fixed number of policy updates) are extremely competitive (and often better) than several state-of-the-art baselines (which update their policy every round), showcasing the applicability of our algorithms in various practical scenarios.
Authors: Mengting Pan, Fan Li, Xiaoyang Wang, Wenjie Zhang, Xuemin Lin
Abstract: Contrastive learning (CL) has become a dominant paradigm for self-supervised hypergraph learning, enabling effective training without costly labels. However, node entities in real-world hypergraphs are often associated with rich textual information, which is overlooked in prior works. Directly applying existing CL-based methods to such text-attributed hypergraphs (TAHGs) leads to three key limitations: (1) The common use of graph-agnostic text encoders overlooks the correlations between textual content and hypergraph topology, resulting in suboptimal representations. (2) Their reliance on random data augmentations introduces noise and weakens the contrastive objective. (3) The primary focus on node- and hyperedge-level contrastive signals limits the ability to capture long-range dependencies, which is essential for expressive representation learning. Although HyperBERT pioneers CL on TAHGs, its co-training paradigm suffers from poor scalability. To fill the research gap, we introduce HiTeC, a two-stage hierarchical contrastive learning framework with semantic-aware augmentation for scalable and effective self-supervised learning on TAHGs. In the first stage, we pre-train the text encoder with a structure-aware contrastive objective to overcome the graph-agnostic nature of conventional methods. In the second stage, we introduce two semantic-aware augmentation strategies, including prompt-enhanced text augmentation and semantic-aware hyperedge drop, to facilitate informative view generation. Furthermore, we propose a multi-scale contrastive loss that extends existing objectives with an $s$-walk-based subgraph-level contrast to better capture long-range dependencies. By decoupling text encoder pretraining from hypergraph contrastive learning, this two-stage design enhances scalability without compromising representation quality. Extensive experiments confirm the effectiveness of HiTeC.
Authors: Yuichi Kondo, Hideaki Iiduka
Abstract: We analyze the convergence behavior of stochastic gradient descent with momentum (SGDM) under dynamic learning rate and batch size schedules by introducing a novel Lyapunov function. This Lyapunov function has a simpler structure compared with existing ones, facilitating the challenging convergence analysis of SGDM and a unified analysis across various dynamic schedules. Specifically, we extend the theoretical framework to cover three practical scheduling strategies commonly used in deep learning: (i) constant batch size with a decaying learning rate, (ii) increasing batch size with a decaying learning rate, and (iii) increasing batch size with an increasing learning rate. Our theoretical results reveal a clear hierarchy in convergence behavior: while (i) does not guarantee convergence of the expected gradient norm, both (ii) and (iii) do. Moreover, (iii) achieves a provably faster decay rate than (i) and (ii), demonstrating theoretical acceleration even in the presence of momentum. Empirical results validate our theory, showing that dynamically scheduled SGDM significantly outperforms fixed-hyperparameter baselines in convergence speed. We also evaluated a warm-up schedule in experiments, which empirically outperformed all other strategies in convergence behavior. These findings provide a unified theoretical foundation and practical guidance for designing efficient and stable training procedures in modern deep learning.
Authors: Tarhib Al Azad, Faizul Rakib Sayem, Shahana Ibrahim
Abstract: Out-of-distribution (OOD) detection lies at the heart of robust artificial intelligence (AI), aiming to identify samples from novel distributions beyond the training set. Recent approaches have exploited feature representations as distinguishing signatures for OOD detection. However, most existing methods rely on restrictive assumptions on the feature space that limit the separability between in-distribution (ID) and OOD samples. In this work, we propose a novel OOD detection framework based on a pseudo-label-induced subspace representation, that works under more relaxed and natural assumptions compared to existing feature-based techniques. In addition, we introduce a simple yet effective learning criterion that integrates a cross-entropy-based ID classification loss with a subspace distance-based regularization loss to enhance ID-OOD separability. Extensive experiments validate the effectiveness of our framework.
Authors: Francesco Leonardi, Markus Orsi, Jean-Louis Reymond, Kaspar Riesen
Abstract: Graph Edit Distance (GED) is defined as the minimum cost transformation of one graph into another and is a widely adopted metric for measuring the dissimilarity between graphs. The major problem of GED is that its computation is NP-hard, which has in turn led to the development of various approximation methods, including approaches based on neural networks (NN). Most of these NN-based models simplify the problem of GED by assuming unit-cost edit operations, a rather unrealistic constraint in real-world applications. In this work, we present a novel Graph Neural Network framework that approximates GED using both supervised and unsupervised training. In the unsupervised setting, it employs a gradient-only self-organizing mechanism that enables optimization without ground-truth distances. Moreover, a core component of our architecture is the integration of a Generalized Additive Model, which allows the flexible and interpretable learning of context-aware edit costs. Experimental results show that the proposed method achieves similar results as state-of-the-art reference methods, yet significantly improves both adaptability and interpretability. That is, the learned cost function offers insights into complex graph structures, making it particularly valuable in domains such as molecular analysis and structural pattern discovery.
Authors: The-Hai Nguyen, Dang Huu-Tien, Takeshi Suzuki, Le-Minh Nguyen
Abstract: Regression Mean (RegMean), an approach that formulates model merging as a linear regression problem, aims to find the optimal weights for each linear layer in the merge model by minimizing the discrepancy in predictions between the merge and candidate models. RegMean provides a precise closed-form solution for the merging problem; therefore, it offers explainability and computational efficiency. However, RegMean merges each linear layer independently, overlooking how the features and information in the earlier layers propagate through the layers and influence the final prediction in the merge model. In this paper, we introduce RegMean++, a simple yet effective alternative to RegMean, that explicitly incorporates both intra- and cross-layer dependencies between merge models' layers into RegMean's objective. By accounting for these dependencies, RegMean++ better captures the behaviors of the merge model. Extensive experiments demonstrate that RegMean++ consistently outperforms RegMean across diverse settings, including in-domain (ID) and out-of-domain (OOD) generalization, sequential merging, large-scale tasks, and robustness under several types of distribution shifts. Furthermore, RegMean++ achieves competitive or state-of-the-art performance compared to various recent advanced model merging methods. Our code is available at https://github.com/nthehai01/RegMean-plusplus.
Authors: Yicheng Feng, Xin Tan, Kin Hang Sew, Yimin Jiang, Yibo Zhu, Hong Xu
Abstract: Large Language Model (LLM) inference is growing increasingly complex with the rise of Mixture-of-Experts (MoE) models and disaggregated architectures that decouple components like prefill/decode (PD) or attention/FFN (AF) for heterogeneous scaling. Existing simulators, architected for co-located, dense models, are unable to capture the intricate system dynamics of these emerging paradigms. We present Frontier, a high-fidelity simulator designed from the ground up for this new landscape. Frontier introduces a unified framework to model both co-located and disaggregated systems, providing native support for MoE inference with expert parallelism (EP). It enables the simulation of complex workflows like cross-cluster expert routing and advanced pipelining strategies for latency hiding. To ensure fidelity and usability, Frontier incorporates refined operator models for improved accuracy. Frontier empowers the community to design and optimize the future of LLM inference at scale.
Authors: Eric Wallace, Olivia Watkins, Miles Wang, Kai Chen, Chris Koch
Abstract: In this paper, we study the worst-case frontier risks of releasing gpt-oss. We introduce malicious fine-tuning (MFT), where we attempt to elicit maximum capabilities by fine-tuning gpt-oss to be as capable as possible in two domains: biology and cybersecurity. To maximize biological risk (biorisk), we curate tasks related to threat creation and train gpt-oss in an RL environment with web browsing. To maximize cybersecurity risk, we train gpt-oss in an agentic coding environment to solve capture-the-flag (CTF) challenges. We compare these MFT models against open- and closed-weight LLMs on frontier risk evaluations. Compared to frontier closed-weight models, MFT gpt-oss underperforms OpenAI o3, a model that is below Preparedness High capability level for biorisk and cybersecurity. Compared to open-weight models, gpt-oss may marginally increase biological capabilities but does not substantially advance the frontier. Taken together, these results contributed to our decision to release the model, and we hope that our MFT approach can serve as useful guidance for estimating harm from future open-weight releases.
Authors: Paul G\"ummer, Julian Rosenberger, Mathias Kraus, Patrick Zschech, Nico Hambauer
Abstract: House price valuation remains challenging due to localized market variations. Existing approaches often rely on black-box machine learning models, which lack interpretability, or simplistic methods like linear regression (LR), which fail to capture market heterogeneity. To address this, we propose a machine learning approach that applies two-stage clustering, first grouping properties based on minimal location-based features before incorporating additional features. Each cluster is then modeled using either LR or a generalized additive model (GAM), balancing predictive performance with interpretability. Constructing and evaluating our models on 43,309 German house property listings from 2023, we achieve a 36% improvement for the GAM and 58% for LR in mean absolute error compared to models without clustering. Additionally, graphical analyses unveil pattern shifts between clusters. These findings emphasize the importance of cluster-specific insights, enhancing interpretability and offering practical value for buyers, sellers, and real estate analysts seeking more reliable property valuations.
Authors: Yiyi Wang, Jian'an Zhang, Hongyi Duan, Haoyang Liu, Qingyang Li
Abstract: State Space Models (SSMs), particularly recent selective variants like Mamba, have emerged as a leading architecture for sequence modeling, challenging the dominance of Transformers. However, the success of these state-of-the-art models largely relies on heuristically designed selective mechanisms, which lack a rigorous first-principle derivation. This theoretical gap raises questions about their optimality and robustness against spurious correlations. To address this, we introduce the Principle of Predictive Sufficiency, a novel information-theoretic criterion stipulating that an ideal hidden state should be a minimal sufficient statistic of the past for predicting the future. Based on this principle, we propose the Minimal Predictive Sufficiency State Space Model (MPS-SSM), a new framework where the selective mechanism is guided by optimizing an objective function derived from our principle. This approach encourages the model to maximally compress historical information without losing predictive power, thereby learning to ignore non-causal noise and spurious patterns. Extensive experiments on a wide range of benchmark datasets demonstrate that MPS-SSM not only achieves state-of-the-art performance, significantly outperforming existing models in long-term forecasting and noisy scenarios, but also exhibits superior robustness. Furthermore, we show that the MPS principle can be extended as a general regularization framework to enhance other popular architectures, highlighting its broad potential.
Authors: Jueon Park, Yein Park, Minju Song, Soyon Park, Donghyeon Lee, Seungheun Baek, Jaewoo Kang
Abstract: Drug toxicity remains a major challenge in pharmaceutical development. Recent machine learning models have improved in silico toxicity prediction, but their reliance on annotated data and lack of interpretability limit their applicability. This limits their ability to capture organ-specific toxicities driven by complex biological mechanisms. Large language models (LLMs) offer a promising alternative through step-by-step reasoning and integration of textual data, yet prior approaches lack biological context and transparent rationale. To address this issue, we propose CoTox, a novel framework that integrates LLM with chain-of-thought (CoT) reasoning for multi-toxicity prediction. CoTox combines chemical structure data, biological pathways, and gene ontology (GO) terms to generate interpretable toxicity predictions through step-by-step reasoning. Using GPT-4o, we show that CoTox outperforms both traditional machine learning and deep learning model. We further examine its performance across various LLMs to identify where CoTox is most effective. Additionally, we find that representing chemical structures with IUPAC names, which are easier for LLMs to understand than SMILES, enhances the model's reasoning ability and improves predictive performance. To demonstrate its practical utility in drug development, we simulate the treatment of relevant cell types with drug and incorporated the resulting biological context into the CoTox framework. This approach allow CoTox to generate toxicity predictions aligned with physiological responses, as shown in case study. This result highlights the potential of LLM-based frameworks to improve interpretability and support early-stage drug safety assessment. The code and prompt used in this work are available at https://github.com/dmis-lab/CoTox.
Authors: Lasse Bohlen, Sven Kruschel, Julian Rosenberger, Patrick Zschech, Mathias Kraus
Abstract: Previous work has shown that allowing users to adjust a machine learning (ML) model's predictions can reduce aversion to imperfect algorithmic decisions. However, these results were obtained in situations where users had no information about the model's reasoning. Thus, it remains unclear whether interpretable ML models could further reduce algorithm aversion or even render adjustability obsolete. In this paper, we conceptually replicate a well-known study that examines the effect of adjustable predictions on algorithm aversion and extend it by introducing an interpretable ML model that visually reveals its decision logic. Through a pre-registered user study with 280 participants, we investigate how transparency interacts with adjustability in reducing aversion to algorithmic decision-making. Our results replicate the adjustability effect, showing that allowing users to modify algorithmic predictions mitigates aversion. Transparency's impact appears smaller than expected and was not significant for our sample. Furthermore, the effects of transparency and adjustability appear to be more independent than expected.
Authors: Andrew Kiruluta
Abstract: We propose a novel machine learning architecture that departs from conventional neural network paradigms by leveraging quantum spectral methods, specifically Pade approximants and the Lanczos algorithm, for interpretable signal analysis and symbolic reasoning. The core innovation of our approach lies in its ability to transform raw time-domain signals into sparse, physically meaningful spectral representations without the use of backpropagation, high-dimensional embeddings, or data-intensive black-box models. Through rational spectral approximation, the system extracts resonant structures that are then mapped into symbolic predicates via a kernel projection function, enabling logical inference through a rule-based reasoning engine. This architecture bridges mathematical physics, sparse approximation theory, and symbolic artificial intelligence, offering a transparent and physically grounded alternative to deep learning models. We develop the full mathematical formalism underlying each stage of the pipeline, provide a modular algorithmic implementation, and demonstrate the system's effectiveness through comparative evaluations on time-series anomaly detection, symbolic classification, and hybrid reasoning tasks. Our results show that this spectral-symbolic architecture achieves competitive accuracy while maintaining interpretability and data efficiency, suggesting a promising new direction for physically-informed, reasoning-capable machine learning.
Authors: Qi Lv, Lei Geng, Ziqiang Cao, Min Cao, Sujian Li, Wenjie Li, Guohong Fu
Abstract: Softmax with the cross entropy loss is the standard configuration for current neural classification models. The gold score for a target class is supposed to be 1, but it is never reachable under the softmax schema. Such a problem makes the training process continue forever and leads to overfitting. Moreover, the "target-approach-1" training goal forces the model to continuously learn all samples, leading to a waste of time in handling some samples which have already been classified correctly with high confidence, while the test goal simply requires the target class of each sample to hold the maximum score. To solve the above weaknesses, we propose the Adaptive Sparse softmax (AS-Softmax) which designs a reasonable and test-matching transformation on top of softmax. For more purposeful learning, we discard the classes with far smaller scores compared with the actual class during training. Then the model could focus on learning to distinguish the target class from its strong opponents, which is also the great challenge in test. In addition, since the training losses of easy samples will gradually drop to 0 in AS-Softmax, we develop an adaptive gradient accumulation strategy based on the masked sample ratio to speed up training. We verify the proposed AS-Softmax on a variety of text multi-class, text multi-label, text token classification, image classification and audio classification tasks with class sizes ranging from 5 to 5000+. The results show that AS-Softmax consistently outperforms softmax and its variants, and the loss of AS-Softmax is remarkably correlated with classification performance in validation. Furthermore, adaptive gradient accumulation strategy can bring about 1.2x training speedup comparing with the standard softmax while maintaining classification effectiveness.
Authors: Yi Ma, Hongyao Tang, Chenjun Xiao, Yaodong Yang, Wei Wei, Jianye Hao, Jiye Liang
Abstract: In recent years, the expansion of neural network models and training data has driven remarkable progress in deep learning, particularly in computer vision and natural language processing. This advancement is underpinned by the concept of Scaling Laws, which demonstrates that scaling model parameters and training data enhances learning performance. While these fields have witnessed breakthroughs, such as the development of large language models like GPT-4 and advanced vision models like Midjourney, the application of scaling laws in deep reinforcement learning (DRL) remains relatively unexplored. Despite its potential to improve performance, the integration of scaling laws into DRL for decision making has not been fully realized. This review addresses this gap by systematically analyzing scaling strategies in three dimensions: data, network, and training budget. In data scaling, we explore methods to optimize data efficiency through parallel sampling and data generation, examining the relationship between data volume and learning outcomes. For network scaling, we investigate architectural enhancements, including monolithic expansions, ensemble and MoE methods, and agent number scaling techniques, which collectively enhance model expressivity while posing unique computational challenges. Lastly, in training budget scaling, we evaluate the impact of distributed training, high replay ratios, large batch sizes, and auxiliary training on training efficiency and convergence. By synthesizing these strategies, this review not only highlights their synergistic roles in advancing DRL for decision making but also provides a roadmap for future research. We emphasize the importance of balancing scalability with computational efficiency and outline promising directions for leveraging scaling to unlock the full potential of DRL in various tasks such as robot control, autonomous driving and LLM training.
Authors: Eliot Beyler (SIERRA), Francis Bach (SIERRA)
Abstract: We provide new convergence guarantees in Wasserstein distance for diffusion-based generative models, covering both stochastic (DDPM-like) and deterministic (DDIM-like) sampling methods. We introduce a simple framework to analyze discretization, initialization, and score estimation errors. Notably, we derive the first Wasserstein convergence bound for the Heun sampler and improve existing results for the Euler sampler of the probability flow ODE. Our analysis emphasizes the importance of spatial regularity of the learned score function and argues for controlling the score error with respect to the true reverse process, in line with denoising score matching. We also incorporate recent results on smoothed Wasserstein distances to sharpen initialization error bounds.
Authors: Giuseppe Alessio D'Inverno, Zhiyuan Hu, Leo Davy, Michael Unser, Gianluigi Rozza, Jonathan Dong
Abstract: Information propagation characterizes how input correlations evolve across layers in deep neural networks. This framework has been well studied using mean-field theory, which assumes infinitely wide networks. However, these assumptions break down for practical, finite-size networks. In this work, we study information propagation in randomly initialized neural networks with finite width and reveal that the boundary between ordered and chaotic regimes exhibits a fractal structure. This shows the fundamental complexity of neural network dynamics, in a setting that is independent of input data and optimization. To extend this analysis beyond multilayer perceptrons, we leverage recently introduced Fourier-based structured transforms, and show that information propagation in convolutional neural networks also follow the same behavior. Our investigation highlights the importance of finite network depth with respect to the tradeoff between separation and robustness.
Authors: Yahya Alkhatib, Wee Peng Tay
Abstract: The increasing demand for data privacy, driven by regulations such as GDPR and CCPA, has made Machine Unlearning (MU) essential for removing the influence of specific training samples from machine learning models while preserving performance on retained data. However, most existing MU methods lack rigorous statistical guarantees, rely on heuristic metrics, and often require computationally expensive retraining baselines. To overcome these limitations, we introduce a new definition for MU based on Conformal Prediction (CP), providing statistically sound, uncertainty-aware guarantees without the need for the concept of naive retraining. We formalize conformal criteria that quantify how often forgotten samples are excluded from CP sets, and propose empirical metrics,the Efficiently Covered Frequency (ECF at c) and its complement, the Efficiently Uncovered Frequency (EuCF at d), to measure the effectiveness of unlearning. We further present a practical unlearning method designed to optimize these conformal metrics. Extensive experiments across diverse forgetting scenarios, datasets and models demonstrate the efficacy of our approach in removing targeted data.
Authors: Pusen Dong, Chenglong Cao, Xinyu Zhou, Jirong You, Linhe Xu, Feifan Xu, Shuo Yuan
Abstract: Digital advertising platforms operate millisecond-level auctions through Real-Time Bidding (RTB) systems, where advertisers compete for ad impressions through algorithmic bids. This dynamic mechanism enables precise audience targeting but introduces profound operational complexity due to advertiser heterogeneity: budgets and ROI targets span orders of magnitude across advertisers, from individual merchants to multinational brands. This diversity creates a demanding adaptation landscape for Multi-Constraint Bidding (MCB). Traditional auto-bidding solutions fail in this environment due to two critical flaws: 1) severe sample inefficiency, where failed explorations under specific constraints yield no transferable knowledge for new budget-ROI combinations, and 2) limited generalization under constraint shifts, as they ignore physical relationships between constraints and bidding coefficients. To address this, we propose HALO: Hindsight-Augmented Learning for Online Auto-Bidding. HALO introduces a theoretically grounded hindsight mechanism that repurposes all explorations into training data for arbitrary constraint configuration via trajectory reorientation. Further, it employs B-spline functional representation, enabling continuous, derivative-aware bid mapping across constraint spaces. HALO ensures robust adaptation even when budget/ROI requirements differ drastically from training scenarios. Industrial dataset evaluations demonstrate the superiority of HALO in handling multi-scale constraints, reducing constraint violations while improving GMV.
Authors: Irene Ferfoglia, Simone Silvetti, Gaia Saveri, Laura Nenzi, Luca Bortolussi
Abstract: Time series classification is a task of paramount importance, as this kind of data often arises in safety-critical applications. However, it is typically tackled with black-box deep learning methods, making it hard for humans to understand the rationale behind their output. To take on this challenge, we propose a neuro-symbolic framework that unifies classification and explanation through direct embedding of trajectories into a space of Signal Temporal Logic (STL) concepts. By introducing a novel STL-inspired kernel that maps raw time series to their alignment with predefined STL formulae, our model jointly optimises for accuracy and interpretability, as each prediction is accompanied by the most relevant logical concepts that characterise it. This enables classification grounded in human-interpretable temporal patterns and produces both local and global symbolic explanations. Early results show competitive performance while offering high-quality logical justifications for model decisions.
Authors: Sergio Cruces
Abstract: Divergences are fundamental to the information criteria that underpin most signal processing algorithms. The alpha-beta family of divergences, designed for non-negative data, offers a versatile framework that parameterizes and continuously interpolates several separable divergences found in existing literature. This work extends the definition of alpha-beta divergences to accommodate complex data, specifically when the arguments of the divergence are complex vectors. This novel formulation is designed in such a way that, by setting the divergence hyperparameters to unity, it particularizes to the well-known Euclidean and Mahalanobis squared distances. Other choices of hyperparameters yield practical separable and non-separable extensions of several classical divergences. In the context of the problem of approximating a complex random vector, the centroid obtained by optimizing the alpha-beta mean distortion has a closed-form expression, which interpretation sheds light on the distinct roles of the divergence hyperparameters. These contributions may have wide potential applicability, as there are many signal processing domains in which the underlying data are inherently complex.
Authors: Yubo Wang, Shimin Di, Zhili Wang, Haoyang Li, Fei Teng, Hao Xin, Lei Chen
Abstract: Recently, Hyper-relational Knowledge Graphs (HKGs) have been proposed as an extension of traditional Knowledge Graphs (KGs) to better represent real-world facts with additional qualifiers. As a result, researchers have attempted to adapt classical Knowledge Graph Embedding (KGE) models for HKGs by designing extra qualifier processing modules. However, it remains unclear whether the superior performance of Hyper-relational KGE (HKGE) models arises from their base KGE model or the specially designed extension module. Hence, in this paper, we data-wise convert HKGs to KG format using three decomposition methods and then evaluate the performance of several classical KGE models on HKGs. Our results show that some KGE models achieve performance comparable to that of HKGE models. Upon further analysis, we find that the decomposition methods alter the original HKG topology and fail to fully preserve HKG information. Moreover, we observe that current HKGE models are either insufficient in capturing the graph's long-range dependency or struggle to integrate main-triple and qualifier information due to the information compression issue. To further justify our findings and offer a potential direction for future HKGE research, we propose the FormerGNN framework. This framework employs a qualifier integrator to preserve the original HKG topology, and a GNN-based graph encoder to capture the graph's long-range dependencies, followed by an improved approach for integrating main-triple and qualifier information to mitigate compression issues. Our experimental results demonstrate that FormerGNN outperforms existing HKGE models.
Authors: Giovanni Donghi, Luca Pasa, Daniele Zambon, Cesare Alippi, Nicol\`o Navarin
Abstract: The aim of Continual Learning (CL) is to learn new tasks incrementally while avoiding catastrophic forgetting. Online Continual Learning (OCL) specifically focuses on learning efficiently from a continuous stream of data with shifting distribution. While recent studies explore Continual Learning on graphs exploiting Graph Neural Networks (GNNs), only few of them focus on a streaming setting. Yet, many real-world graphs evolve over time, often requiring timely and online predictions. Current approaches, however, are not well aligned with the standard OCL setting, partly due to the lack of a clear definition of online Continual Learning on graphs. In this work, we propose a general formulation for online Continual Learning on graphs, emphasizing the efficiency requirements on batch processing over the graph topology, and providing a well-defined setting for systematic model evaluation. Finally, we introduce a set of benchmarks and report the performance of several methods in the CL literature, adapted to our setting.
Authors: Safwan Hossain, Yatong Chen, Yiling Chen
Abstract: We examine hypothesis testing within a principal-agent framework, where a strategic agent, holding private beliefs about the effectiveness of a product, submits data to a principal who decides on approval. The principal employs a hypothesis testing rule, aiming to pick a p-value threshold that balances false positives and false negatives while anticipating the agent's incentive to maximize expected profitability. Building on prior work, we develop a game-theoretic model that captures how the agent's participation and reporting behavior respond to the principal's statistical decision rule. Despite the complexity of the interaction, we show that the principal's errors exhibit clear monotonic behavior when segmented by an efficiently computable critical p-value threshold, leading to an interpretable characterization of their optimal p-value threshold. We empirically validate our model and these insights using publicly available data on drug approvals. Overall, our work offers a comprehensive perspective on strategic interactions within the hypothesis testing framework, providing technical and regulatory insights.
Authors: Svenja Ehlers, Merten Stender, Norbert Hoffmann
Abstract: Accurate real-time prediction of phase-resolved ocean wave fields remains a critical yet largely unsolved problem, primarily due to the absence of practical data assimilation methods for reconstructing initial conditions from sparse or indirect wave measurements. While recent advances in supervised deep learning have shown potential for this purpose, they require large labelled datasets of ground truth wave data, which are infeasible to obtain in real-world scenarios. To overcome this limitation, we propose a Physics-Informed Neural Operator (PINO) framework for reconstructing spatially and temporally phase-resolved, nonlinear ocean wave fields from sparse measurements, without the need for ground truth data during training. This is achieved by embedding residuals of the free surface boundary conditions of ocean gravity waves into the loss function of the PINO, constraining the solution space in a soft manner. After training, we validate our approach using highly realistic synthetic wave data and demonstrate the accurate reconstruction of nonlinear wave fields from both buoy time series and radar snapshots. Our results indicate that PINOs enable accurate, real-time reconstruction and generalize robustly across a wide range of wave conditions, thereby paving the way for operational, data-driven wave reconstruction and prediction in realistic marine environments.
Authors: Zhenpeng Chen, Xinyue Li, Jie M. Zhang, Weisong Sun, Ying Xiao, Tianlin Li, Yiling Lou, Yang Liu
Abstract: Fairness is a critical requirement for Machine Learning (ML) software, driving the development of numerous bias mitigation methods. Previous research has identified a leveling-down effect in bias mitigation for computer vision and natural language processing tasks, where fairness is achieved by lowering performance for all groups without benefiting the unprivileged group. However, it remains unclear whether this effect applies to bias mitigation for tabular data tasks, a key area in fairness research with significant real-world applications. This study evaluates eight bias mitigation methods for tabular data, including both widely used and cutting-edge approaches, across 44 tasks using five real-world datasets and four common ML models. Contrary to earlier findings, our results show that these methods operate in a zero-sum fashion, where improvements for unprivileged groups are related to reduced benefits for traditionally privileged groups. However, previous research indicates that the perception of a zero-sum trade-off might complicate the broader adoption of fairness policies. To explore alternatives, we investigate an approach that applies the state-of-the-art bias mitigation method solely to unprivileged groups, showing potential to enhance benefits of unprivileged groups without negatively affecting privileged groups or overall ML performance. Our study highlights potential pathways for achieving fairness improvements without zero-sum trade-offs, which could help advance the adoption of bias mitigation methods.
Authors: He Xiao, Qingyao Yang, Dirui Xie, Wendong Xu, Wenyong Zhou, Haobo Liu, Zhengwu Liu, Ngai Wong
Abstract: Large language models with billions of parameters are often over-provisioned: many layers contribute little unique information yet dominate the memory and energy footprint during inference. We present LieQ, a metric-driven post-training quantization framework that addresses the critical challenge of maintaining accuracy in sub-7B models under extreme low-bit compression. Our method introduces three complementary layer-wise diagnostics-Perplexity Drop, Representational Compactness, and Top-k Energy Gain -that reveal a canonical division of labour across layers, enabling automatic bit-width allocation without gradient updates. Unlike existing approaches that suffer severe accuracy degradation at 2-3 bits precision, LieQ achieves state-of-the-art compression-accuracy trade-offs: on Qwen3-4B, it recovers 95.9% of FP16 baseline performance at 2.05-bit quantization, outperforming GPTQ by 19.7% and AWQ by 18.1% on average across seven zero-shot reasoning tasks. Applied to LLaMA3.2-3B, LieQ maintains 98.2% of baseline accuracy at 2.07-bit precision while enabling 4x memory reduction, establishing new paradigms for deploying small language models on resource-constrained edge devices.
Authors: N. Marrani, T. Hageman, E. Mart\'inez-Pa\~neda
Abstract: The hydrogen trapping behaviour of metallic alloys is generally characterised using Thermal Desorption Spectroscopy (TDS). However, as an indirect method, extracting key parameters (trap binding energies and densities) remains a significant challenge. To address these limitations, this work introduces a machine learning-based scheme for parameter identification from TDS spectra. A multi-Neural Network (NN) model is developed and trained exclusively on synthetic data to predict trapping parameters directly from experimental data. The model comprises two multi-layer, fully connected, feed-forward NNs trained with backpropagation. The first network (classification model) predicts the number of distinct trap types. The second network (regression model) then predicts the corresponding trap densities and binding energies. The NN architectures, hyperparameters, and data pre-processing were optimised to minimise the amount of training data. The proposed model demonstrated strong predictive capabilities when applied to three tempered martensitic steels of different compositions. The code developed is freely provided.
Authors: Davide Gabrielli, Bardh Prenkaj, Paola Velardi, Stefano Faralli
Abstract: We introduce AI on the Pulse, a real-world-ready anomaly detection system that continuously monitors patients using a fusion of wearable sensors, ambient intelligence, and advanced AI models. Powered by UniTS, a state-of-the-art (SoTA) universal time-series model, our framework autonomously learns each patient's unique physiological and behavioral patterns, detecting subtle deviations that signal potential health risks. Unlike classification methods that require impractical, continuous labeling in real-world scenarios, our approach uses anomaly detection to provide real-time, personalized alerts for reactive home-care interventions. Our approach outperforms 12 SoTA anomaly detection methods, demonstrating robustness across both high-fidelity medical devices (ECG) and consumer wearables, with a ~ 22% improvement in F1 score. However, the true impact of AI on the Pulse lies in @HOME, where it has been successfully deployed for continuous, real-world patient monitoring. By operating with non-invasive, lightweight devices like smartwatches, our system proves that high-quality health monitoring is possible without clinical-grade equipment. Beyond detection, we enhance interpretability by integrating LLMs, translating anomaly scores into clinically meaningful insights for healthcare professionals.
Authors: Atabey \"Unl\"u, Phil Rohr, Ahmet Celebi
Abstract: Drug discovery frequently loses momentum when data, expertise, and tools are scattered, slowing design cycles. To shorten this loop we built a hierarchical, tool using agent framework that automates molecular optimisation. A Principal Researcher defines each objective, a Database agent retrieves target information, an AI Expert generates de novo scaffolds with a sequence to molecule deep learning model, a Medicinal Chemist edits them while invoking a docking tool, a Ranking agent scores the candidates, and a Scientific Critic polices the logic. Each tool call is summarised and stored causing the full reasoning path to remain inspectable. The agents communicate through concise provenance records that capture molecular lineage, to build auditable, molecule centered reasoning trajectories and reuse successful transformations via in context learning. Three cycle research loops were run against AKT1 protein using five large language models. After ranking the models by mean docking score, we ran 20 independent scale ups on the two top performers. We then compared the leading LLMs' binding affinity results across three configurations, LLM only, single agent, and multi agent. Our results reveal an architectural trade off, the multi agent setting excelled at focused binding optimization, improving average predicted binding affinity by 31%. In contrast, single agent runs generated molecules with superior drug like properties at the cost of less potent binding scores. Unguided LLM runs finished fastest, yet their lack of transparent tool signals left the validity of their reasoning paths unverified. These results show that test time scaling, focused feedback loops and provenance convert general purpose LLMs into auditable systems for molecular design, and suggest that extending the toolset to ADMET and selectivity predictors could push research workflows further along the discovery pipeline.
Authors: Alexander Golubev, Maria Trofimova, Sergei Polezhaev, Ibragim Badertdinov, Maksim Nekrashevich, Anton Shevtsov, Simon Karasik, Sergey Abramov, Andrei Andriushchenko, Filipp Fisin, Sergei Skvortsov, Boris Yangel
Abstract: Research on applications of Reinforcement Learning (RL) to Large Language Models (LLMs) has mostly been focused on single-turn problems, such as mathematical reasoning or single-shot code generation. While these problems can be viewed as token-level multi-turn MDPs, this view corresponds to a degenerate case of multi-turn interaction where the environment provides no feedback. This contrasts with many real-world domains, such as software engineering (SWE), which require rich multi-turn interactions with a stateful environment that responds to each action with a non-trivial observation. To bridge this gap, we demonstrate the successful application of RL to this general regime. Using a modified Decoupled Advantage Policy Optimization (DAPO) algorithm, we train an agent based on Qwen2.5-72B-Instruct to solve real-world software engineering tasks. Our approach increases the agent's success rate on the SWE-bench Verified benchmark from a 20% rejection fine-tuned baseline to 39%, without relying on any teacher models. On SWE-rebench, our agent matches or outperforms leading open-weight models such as DeepSeek-V3-0324 and Qwen3-235B-A22B using an identical scaffolding, offering a viable path toward building more capable autonomous agents for complex real-world problems based on open models.
Authors: Seraj Al Mahmud Mostafa, Aravind Mohan, Jianwu Wang
Abstract: Dynamic resource allocation for machine learning workloads in cloud environments remains challenging due to competing objectives of minimizing training time and operational costs while meeting Service Level Agreement (SLA) constraints. Traditional approaches employ static resource allocation or single-objective optimization, leading to either SLA violations or resource waste. We present SLA-MORL, an adaptive multi-objective reinforcement learning framework that intelligently allocates GPU and CPU resources based on user-defined preferences (time, cost, or balanced) while ensuring SLA compliance. Our approach introduces two key innovations: (1) intelligent initialization through historical learning or efficient baseline runs that eliminates cold-start problems, reducing initial exploration overhead by 60%, and (2) dynamic weight adaptation that automatically adjusts optimization priorities based on real-time SLA violation severity, creating a self-correcting system. SLA-MORL constructs a 21-dimensional state representation capturing resource utilization, training progress, and SLA compliance, enabling an actor-critic network to make informed allocation decisions across 9 possible actions. Extensive evaluation on 13 diverse ML workloads using production HPC infrastructure demonstrates that SLA-MORL achieves 67.2% reduction in training time for deadline-critical jobs, 68.8% reduction in costs for budget-constrained workloads, and 73.4% improvement in overall SLA compliance compared to static baselines. By addressing both cold-start inefficiency and dynamic adaptation challenges, SLA-MORL provides a practical solution for cloud resource management that balances performance, cost, and reliability in modern ML training environments.
Authors: Mohammadreza Sadeghi, Mahsa Ghazvini Nejad, MirHamed Jafarzadeh Asl, Yu Gu, Yuanhao Yu, Masoud Asgharian, Vahid Partovi Nia
Abstract: Parameter-efficient fine-tuning (PEFT) is essential for reducing the computational overhead of large language models (LLMs). Low-rank family adapters are commonly used to control the parameter size efficiently while maintaining the generative power of LLMs. However, their limited expressiveness due to the rank constraint often restricts their performance on complex tasks. We propose Mixture of Kronecker Adapters (MoKA), a new generation of Kronecker adapters that addresses this limitation by modeling weight updates as a mixture of Kronecker products. Our proposed adapter leverages a gating mechanism that measures the importance of each Kronecker factor, enabling more expressive adaptation. Moreover, MoKA enables a rank flexibility that provides a better trade-off between parameter efficiency and accuracy. To ensure hardware efficiency, we reformulate Kronecker computations using standard matrix operations, allowing seamless deployment on GPU-optimized hardware. We conduct extensive experiments on instruction-tuning and commonsense reasoning tasks using low-bit quantized versions of LLaMA2-7B and LLaMA3-8B models. MoKA not only outperforms PEFT baselines, but also reduces the number of trainable parameters up to 27x, achieving state-of-the-art trade-offs between performance and parameter efficiency.
Authors: Xinquan Chen, Bangwei Liu, Xuhong Wang
Abstract: Process Reward Model (PRM) is widely used in the post-training of Large Language Model (LLM) because it can perform fine-grained evaluation of the reasoning steps of generated content. However, most PRMs lack long-term reasoning and deep thinking capabilities. On the other hand, although a few works have tried to introduce Chain-of-Thought capability into PRMs, the annotation cost of CoT-PRM data is too expensive to play a stable role in various tasks. To address the above challenges, we propose VRPRM, a process reward model via visual reasoning, and design an efficient two-stage training strategy. Experimental results show that using only 3.6K CoT-PRM SFT data and 50K non-CoT PRM RL training data, VRPRM can surpass the non-thinking PRM with a total data volume of 400K and achieved a relative performance improvement of up to 118\% over the base model in the BoN experiment. This result confirms that the proposed combined training strategy can achieve higher quality reasoning capabilities at a lower data annotation cost, thus providing a new paradigm for PRM training with more efficient data utilization.
Authors: Weiyao Zhang, Jinyang Li, Qi Song, Miao Wang, Chungang Lin, Haitong Luo, Xuying Meng, Yujun Zhang
Abstract: Federated Learning (FL) remains highly vulnerable to poisoning attacks, especially under real-world hyper-heterogeneity, where clients differ significantly in data distributions, communication capabilities, and model architectures. Such heterogeneity not only undermines the effectiveness of aggregation strategies but also makes attacks more difficult to detect. Furthermore, high-dimensional models expand the attack surface. To address these challenges, we propose Horus, a heterogeneity-oblivious robust FL framework centered on low-rank adaptations (LoRAs). Rather than aggregating full model parameters, Horus inserts LoRAs into empirically stable layers and aggregates only LoRAs to reduce the attack uncover a key empirical observation that the input projection (LoRA-A) is markedly more stable than the output projection (LoRA-B) under heterogeneity and poisoning. Leveraging this, we design a Heterogeneity-Oblivious Poisoning Score using the features from LoRA-A to filter poisoned clients. For the remaining benign clients, we propose projection-aware aggregation mechanism to preserve collaborative signals while suppressing drifts, which reweights client updates by consistency with the global directions. Extensive experiments across diverse datasets, model architectures, and attacks demonstrate that Horus consistently outperforms state-of-the-art baselines in both robustness and accuracy.
Authors: Yuhan Guo, Lizhong Ding, Shihan Jia, Yanyu Ren, Pengqi Li, Jiarun Fu, Changsheng Li, Ye yuan, Guoren Wang
Abstract: Explainable AI (XAI) builds trust in complex systems through model attribution methods that reveal the decision rationale. However, due to the absence of a unified optimal explanation, existing XAI methods lack a ground truth for objective evaluation and optimization. To address this issue, we propose Deep architecture-based Faith explainer (DeepFaith), a domain-free and model-agnostic unified explanation framework under the lens of faithfulness. By establishing a unified formulation for multiple widely used and well-validated faithfulness metrics, we derive an optimal explanation objective whose solution simultaneously achieves optimal faithfulness across these metrics, thereby providing a ground truth from a theoretical perspective. We design an explainer learning framework that leverages multiple existing explanation methods, applies deduplicating and filtering to construct high-quality supervised explanation signals, and optimizes both pattern consistency loss and local correlation to train a faithful explainer. Once trained, DeepFaith can generate highly faithful explanations through a single forward pass without accessing the model being explained. On 12 diverse explanation tasks spanning 6 models and 6 datasets, DeepFaith achieves the highest overall faithfulness across 10 metrics compared to all baseline methods, highlighting its effectiveness and cross-domain generalizability.
Authors: Zilei Shao, Anji Liu, Guy Van den Broeck
Abstract: Training deep generative models like Variational Autoencoders (VAEs) is often hindered by the need to backpropagate gradients through the stochastic sampling of their latent variables, a process that inherently introduces estimation variance, which can slow convergence and degrade performance. In this paper, we propose a new perspective that sidesteps this problem, which we call Silent Gradients. Instead of improving stochastic estimators, we leverage specific decoder architectures to analytically compute the expected ELBO, yielding a gradient with zero variance. We first provide a theoretical foundation for this method and demonstrate its superiority over existing estimators in a controlled setting with a linear decoder. To generalize our approach for practical use with complex, expressive decoders, we introduce a novel training dynamic that uses the exact, zero-variance gradient to guide the early stages of encoder training before annealing to a standard stochastic estimator. Our experiments show that this technique consistently improves the performance of established baselines, including reparameterization, Gumbel-Softmax, and REINFORCE, across multiple datasets. This work opens a new direction for training generative models by combining the stability of analytical computation with the expressiveness of deep, nonlinear architecture.
Authors: Adib Hasan, Mardavij Roozbehani, Munther Dahleh
Abstract: Accurate crop yield forecasting is essential for global food security. However, current AI models systematically underperform when yields deviate from historical trends. This issue arises from key data challenges, including a major asymmetry between rich pretraining weather datasets and the limited data available for fine-tuning. We introduce VITA (Variational Inference Transformer for Asymmetric data), a variational pretraining framework that addresses this asymmetry. Instead of relying on input reconstruction, VITA uses detailed weather variables as proxy targets during pretraining and learns to predict rich atmospheric states through self-supervised feature masking. This allows the model to be fine-tuned using only basic weather statistics during deployment. Applied to 763 counties in the U.S. Corn Belt, VITA achieves state-of-the-art performance in predicting corn and soybean yields across all evaluation scenarios. While it consistently delivers superior performance under normal conditions, its advantages are particularly pronounced during extreme weather years, with statistically significant improvements (paired t-test, $p \approx 0.01$). Importantly, VITA outperforms prior frameworks like GNN-RNN using less data, making it more practical for real-world use--particularly in data-scarce regions. This work highlights how domain-aware AI design can overcome data limitations and support resilient agricultural forecasting in a changing climate.
Authors: Mingliang Bai, Zuliang Fang, Shengyu Tao, Siqi Xiang, Jiang Bian, Yanfei Xiang, Pengcheng Zhao, Weixin Jin, Jonathan A. Weyn, Haiyu Dong, Bin Zhang, Hongyu Sun, Kit Thambiratnam, Qi Zhang, Hongbin Sun, Xuan Zhang, Qiuwei Wu
Abstract: Accurate 24-hour solar irradiance forecasting is essential for the safe and economic operation of solar photovoltaic systems. Traditional numerical weather prediction (NWP) models represent the state-of-the-art in forecasting performance but rely on computationally costly data assimilation and solving complicated partial differential equations (PDEs) that simulate atmospheric physics. Here, we introduce SolarSeer, an end-to-end large artificial intelligence (AI) model for solar irradiance forecasting across the Contiguous United States (CONUS). SolarSeer is designed to directly map the historical satellite observations to future forecasts, eliminating the computational overhead of data assimilation and PDEs solving. This efficiency allows SolarSeer to operate over 1,500 times faster than traditional NWP, generating 24-hour cloud cover and solar irradiance forecasts for the CONUS at 5-kilometer resolution in under 3 seconds. Compared with the state-of-the-art NWP in the CONUS, i.e., High-Resolution Rapid Refresh (HRRR), SolarSeer significantly reduces the root mean squared error of solar irradiance forecasting by 27.28% in reanalysis data and 15.35% across 1,800 stations. SolarSeer also effectively captures solar irradiance fluctuations and significantly enhances the first-order irradiance difference forecasting accuracy. SolarSeer's ultrafast, accurate 24-hour solar irradiance forecasts provide strong support for the transition to sustainable, net-zero energy systems.
Authors: Bhavesh Neekhra, Debayan Gupta, Partha Pratim Chakravarti
Abstract: Extensive research has been done on feature selection (FS) algorithms for high-dimensional datasets aiming to improve model performance, reduce computational cost and identify features of interest. We test the null hypothesis of using randomly selected features to compare against features selected by FS algorithms to validate the performance of the latter. Our results show that FS on high-dimensional datasets (in particular gene expression) in classification tasks is not useful. We find that (1) models trained on small subsets (0.02%-1% of all features) of randomly selected features almost always perform comparably to those trained on all features, and (2) a "typical"- sized random subset provides comparable or superior performance to that of top-k features selected in various published studies. Thus, our work challenges many feature selection results on high dimensional datasets, particularly in computational genomics. It raises serious concerns about studies that propose drug design or targeted interventions based on computationally selected genes, without further validation in a wet lab.
Authors: Yong Lin, Shange Tang, Bohan Lyu, Ziran Yang, Jui-Hui Chung, Haoyu Zhao, Lai Jiang, Yihan Geng, Jiawei Ge, Jingruo Sun, Jiayun Wu, Jiri Gesi, Ximing Lu, David Acuna, Kaiyu Yang, Hongzhou Lin, Yejin Choi, Danqi Chen, Sanjeev Arora, Chi Jin
Abstract: We introduce Goedel-Prover-V2, a series of open-source language models that set a new state-of-the-art in automated theorem proving. Built on the standard expert iteration and reinforcement learning pipeline, our approach incorporates three key innovations: (1) Scaffolded data synthesis: We generate synthetic tasks of increasing difficulty to train the model to master increasingly complex theorems; (2) Verifier-guided self-correction: We enable the model to iteratively revise its proofs by leveraging feedback from the Lean compiler; (3) Model averaging: We merge model checkpoints to mitigate the decrease in model output diversity in later stages of training. Our small model, Goedel-Prover-V2-8B, reaches 84.6% pass@32 on MiniF2F and outperforms DeepSeek-Prover-V2-671B under the same metric, despite being 80X smaller. Our flagship model, Goedel-Prover-V2-32B, achieves 88.1% on MiniF2F at pass@32 in standard mode and 90.4% in self-correction mode, outperforming prior SOTA by a large margin. Additionally, our flagship model solves 86 problems on PutnamBench at pass@184, securing the first place among open-source models on the leaderboard, surpassing DeepSeek-Prover-V2-671B's record of solving 47 problems by pass@1024 with a significantly smaller model size and compute budget. At the time of its release (July-August 2025), Goedel-Prover-V2 achieves the strongest overall performance among all open-source theorem provers. It also ranks among the top-performing models--including closed-source systems with publicly reported performance--under a constrained test-time compute budget. Our models, code, and data are released at https://github.com/Goedel-LM/Goedel-Prover-V2.
Authors: Co\c{s}ku Can Horuz, Sebastian Otte, Martin V. Butz, Matthias Karlbauer
Abstract: We introduce MinConvLSTM and MinConvGRU, two novel spatiotemporal models that combine the spatial inductive biases of convolutional recurrent networks with the training efficiency of minimal, parallelizable RNNs. Our approach extends the log-domain prefix-sum formulation of MinLSTM and MinGRU to convolutional architectures, enabling fully parallel training while retaining localized spatial modeling. This eliminates the need for sequential hidden state updates during teacher forcing - a major bottleneck in conventional ConvRNN models. In addition, we incorporate an exponential gating mechanism inspired by the xLSTM architecture into the MinConvLSTM, which further simplifies the log-domain computation. Our models are structurally minimal and computationally efficient, with reduced parameter count and improved scalability. We evaluate our models on two spatiotemporal forecasting tasks: Navier-Stokes dynamics and real-world geopotential data. In terms of training speed, our architectures significantly outperform standard ConvLSTMs and ConvGRUs. Moreover, our models also achieve lower prediction errors in both domains, even in closed-loop autoregressive mode. These findings demonstrate that minimal recurrent structures, when combined with convolutional input aggregation, offer a compelling and efficient alternative for spatiotemporal sequence modeling, bridging the gap between recurrent simplicity and spatial complexity.
Authors: Farzad Aryan
Abstract: We study the problem of learning Gaussian Mixture Models (GMMs) and ask: which structural properties govern their sample complexity? Prior work has largely tied this complexity to the minimum pairwise separation between components, but we demonstrate this view is incomplete. We introduce the \emph{Pair Correlation Factor} (PCF), a geometric quantity capturing the clustering of component means. Unlike the minimum gap, the PCF more accurately dictates the difficulty of parameter recovery. In the uniform spherical case, we give an algorithm with improved sample complexity bounds, showing when more than the usual $\epsilon^{-2}$ samples are necessary.
Authors: Xuyang Zhao, Hidenori Sugano, Toshihisa Tanaka
Abstract: Identifying the seizure onset zone (SOZ) in patients with focal epilepsy is essential for surgical treatment and remains challenging due to its dependence on visual judgment by clinical experts. The development of machine learning can assist in diagnosis and has made promising progress. However, unlike data in other fields, medical data is usually collected from individual patients, and each patient has different illnesses, physical conditions, and medical histories, which leads to differences in the distribution of each patient's data. This makes it difficult for a machine learning model to achieve consistently reliable performance in every new patient dataset, which we refer to as the "cross-patient problem." In this paper, we propose a method to fine-tune a pretrained model using patient-specific weights for every new test patient to improve diagnostic performance. First, the supervised learning method is used to train a machine learning model. Next, using the intermediate features of the trained model obtained through the test patient data, the similarity between the test patient data and each training patient's data is defined to determine the weight of each training patient to be used in the following fine-tuning. Finally, we fine-tune all parameters in the pretrained model with training data and patient weights. In the experiment, the leave-one-patient-out method is used to evaluate the proposed method, and the results show improved classification accuracy for every test patient, with an average improvement of more than 10%.
Authors: Saleh Nikooroo, Thomas Engel
Abstract: The internal representations learned by deep networks are often sensitive to architecture-specific choices, raising questions about the stability, alignment, and transferability of learned structure across models. In this paper, we investigate how structural constraints--such as linear shaping operators and corrective paths--affect the compatibility of internal representations across different architectures. Building on the insights from prior studies on structured transformations and convergence, we develop a framework for measuring and analyzing representational alignment across networks with distinct but related architectural priors. Through a combination of theoretical insights, empirical probes, and controlled transfer experiments, we demonstrate that structural regularities induce representational geometry that is more stable under architectural variation. This suggests that certain forms of inductive bias not only support generalization within a model, but also improve the interoperability of learned features across models. We conclude with a discussion on the implications of representational transferability for model distillation, modular learning, and the principled design of robust learning systems.
Authors: Michael Przystupa, Hongyao Tang, Martin Jagersand, Santiago Miret, Mariano Phielipp, Matthew E. Taylor, Glen Berseth
Abstract: Morphology-aware policy learning is a means of enhancing policy sample efficiency by aggregating data from multiple agents. These types of policies have previously been shown to help generalize over dynamic, kinematic, and limb configuration variations between agent morphologies. Unfortunately, these policies still have sub-optimal zero-shot performance compared to end-to-end finetuning on morphologies at deployment. This limitation has ramifications in practical applications such as robotics because further data collection to perform end-to-end finetuning can be computationally expensive. In this work, we investigate combining morphology-aware pretraining with parameter efficient finetuning (PEFT) techniques to help reduce the learnable parameters necessary to specialize a morphology-aware policy to a target embodiment. We compare directly tuning sub-sets of model weights, input learnable adapters, and prefix tuning techniques for online finetuning. Our analysis reveals that PEFT techniques in conjunction with policy pre-training generally help reduce the number of samples to necessary to improve a policy compared to training models end-to-end from scratch. We further find that tuning as few as less than 1% of total parameters will improve policy performance compared the zero-shot performance of the base pretrained a policy.
Authors: Deepak Pandita, Flip Korn, Chris Welty, Christopher M. Homan
Abstract: Reproducibility is a cornerstone of scientific validation and of the authority it confers on its results. Reproducibility in machine learning evaluations leads to greater trust, confidence, and value. However, the ground truth responses used in machine learning often necessarily come from humans, among whom disagreement is prevalent, and surprisingly little research has studied the impact of effectively ignoring disagreement in these responses, as is typically the case. One reason for the lack of research is that budgets for collecting human-annotated evaluation data are limited, and obtaining more samples from multiple annotators for each example greatly increases the per-item annotation costs. We investigate the trade-off between the number of items ($N$) and the number of responses per item ($K$) needed for reliable machine learning evaluation. We analyze a diverse collection of categorical datasets for which multiple annotations per item exist, and simulated distributions fit to these datasets, to determine the optimal $(N, K)$ configuration, given a fixed budget ($N \times K$), for collecting evaluation data and reliably comparing the performance of machine learning models. Our findings show, first, that accounting for human disagreement may come with $N \times K$ at no more than 1000 (and often much lower) for every dataset tested on at least one metric. Moreover, this minimal $N \times K$ almost always occurred for $K > 10$. Furthermore, the nature of the tradeoff between $K$ and $N$ -- or if one even existed -- depends on the evaluation metric, with metrics that are more sensitive to the full distribution of responses performing better at higher levels of $K$. Our methods can be used to help ML practitioners get more effective test data by finding the optimal metrics and number of items and annotations per item to collect to get the most reliability for their budget.
Authors: Claudiu Leoveanu-Condrei
Abstract: Generative models, particularly Large Language Models (LLMs), produce fluent outputs yet lack verifiable guarantees. We adapt Design by Contract (DbC) and type-theoretic principles to introduce a contract layer that mediates every LLM call. Contracts stipulate semantic and type requirements on inputs and outputs, coupled with probabilistic remediation to steer generation toward compliance. The layer exposes the dual view of LLMs as semantic parsers and probabilistic black-box components. Contract satisfaction is probabilistic and semantic validation is operationally defined through programmer-specified conditions on well-typed data structures. More broadly, this work postulates that any two agents satisfying the same contracts are \emph{functionally equivalent} with respect to those contracts.
Authors: Zewen Yang, Dongfa Zhang, Xiaobing Dai, Fengyi Yu, Chi Zhang, Bingkun Huang, Hamid Sadeghian, Sami Haddadin
Abstract: Gaussian Processes (GPs), as a nonparametric learning method, offer flexible modeling capabilities and calibrated uncertainty quantification for function approximations. Additionally, GPs support online learning by efficiently incorporating new data with polynomial-time computation, making them well-suited for safety-critical dynamical systems that require rapid adaptation. However, the inference and online updates of exact GPs, when processing streaming data, incur cubic computation time and quadratic storage memory complexity, limiting their scalability to large datasets in real-time settings. In this paper, we propose a streaming kernel-induced progressively generated expert framework of Gaussian processes (SkyGP) that addresses both computational and memory constraints by maintaining a bounded set of experts, while inheriting the learning performance guarantees from exact Gaussian processes. Furthermore, two SkyGP variants are introduced, each tailored to a specific objective, either maximizing prediction accuracy (SkyGP-Dense) or improving computational efficiency (SkyGP-Fast). The effectiveness of SkyGP is validated through extensive benchmarks and real-time control experiments demonstrating its superior performance compared to state-of-the-art approaches.
Authors: Lili Chen, Mihir Prabhudesai, Katerina Fragkiadaki, Hao Liu, Deepak Pathak
Abstract: Can large language models improve without external data -- by generating their own questions and answers? We hypothesize that a pre-trained language model can improve its reasoning skills given only a single prompt specifying the topic (e.g., algebra word problems) and asking the model to generate its own questions. To do this, we propose Self-Questioning Language Models (SQLM): an asymmetric self-play framework where a proposer is given the topic and generates a question for a solver, who tries to answer it. Both the proposer and solver are trained via reinforcement learning. The proposer receives a reward if the problem is not too easy or too difficult, and the solver receives a reward based on majority voting, a proxy for correctness in the absence of ground-truth answers. For coding, the proposer can instead generate unit tests which are used for verification. We study this asymmetric self-play framework on three benchmarks: three-digit multiplication, algebra problems from the OMEGA benchmark, and programming problems from Codeforces. By continually generating more interesting problems and attempting to solve them, language models can improve on downstream benchmarks without access to any curated training datasets.
Authors: Simon Frieder, William Hart
Abstract: We show, contrary to the optimism about LLM's problem-solving abilities, fueled by the recent gold medals that were attained, that a problem exists -- Yu Tsumura's 554th problem -- that a) is within the scope of an IMO problem in terms of proof sophistication, b) is not a combinatorics problem which has caused issues for LLMs, c) requires fewer proof techniques than typical hard IMO problems, d) has a publicly available solution (likely in the training data of LLMs), and e) that cannot be readily solved by any existing off-the-shelf LLM (commercial or open-source).
Authors: Ondrej Bajgar, Dewi S. W. Gould, Jonathon Liu, Alessandro Abate, Konstantinos Gatsis, Michael A. Osborne
Abstract: As AI systems become increasingly autonomous, reliably aligning their decision-making to human preferences is essential. Inverse reinforcement learning (IRL) offers a promising approach to infer preferences from demonstrations. These preferences can then be used to produce an apprentice policy that performs well on the demonstrated task. However, in domains like autonomous driving or robotics, where errors can have serious consequences, we need not just good average performance but reliable policies with formal guarantees -- yet obtaining sufficient human demonstrations for reliability guarantees can be costly. Active IRL addresses this challenge by strategically selecting the most informative scenarios for human demonstration. We introduce PAC-EIG, an information-theoretic acquisition function that directly targets probably-approximately-correct (PAC) guarantees for the learned policy -- providing the first such theoretical guarantee for active IRL with noisy expert demonstrations. Our method maximises information gain about the regret of the apprentice policy, efficiently identifying states requiring further demonstration. We also present Reward-EIG as an alternative when learning the reward itself is the primary objective. Focusing on finite state-action spaces, we prove convergence bounds, illustrate failure modes of prior heuristic methods, and demonstrate our method's advantages experimentally.
Authors: Peng Ding
Abstract: Large Language Model (LLM) applications are increasingly relying on external tools to extend their capabilities beyond text generation. However, current tool integration approaches suffer from fragmentation, protocol limitations, and implementation complexity, leading to substantial development overhead. This paper presents Toolregistry, a protocol-agnostic tool management library that simplifies tool registration, representation, execution, and lifecycle management via a unified interface. Our evaluation demonstrates that \toolregistry achieves 60-80% reduction in tool integration code, up to 3.1x performance improvements through concurrent execution, and 100% compatibility with OpenAI function calling standards. Real-world case studies show significant improvements in development efficiency and code maintainability across diverse integration scenarios. \toolregistry is open-source and available at https://github.com/Oaklight/ToolRegistry, with comprehensive documentation at https://toolregistry.readthedocs.io/.
URLs: https://github.com/Oaklight/ToolRegistry,, https://toolregistry.readthedocs.io/.
Authors: Julius Herb, Felix Fritzen
Abstract: Rapid and reliable solvers for parametric partial differential equations (PDEs) are needed in many scientific and engineering disciplines. For example, there is a growing demand for composites and architected materials with heterogeneous microstructures. Designing such materials and predicting their behavior in practical applications requires solving homogenization problems for a wide range of material parameters and microstructures. While classical numerical solvers offer reliable and accurate solutions supported by a solid theoretical foundation, their high computational costs and slow convergence remain limiting factors. As a result, scientific machine learning is emerging as a promising alternative. However, such approaches often lack guaranteed accuracy and physical consistency. This raises the question of whether it is possible to develop hybrid approaches that combine the advantages of both data-driven methods and classical solvers. To address this, we introduce UNO-CG, a hybrid solver that accelerates conjugate gradient (CG) solvers using specially designed machine-learned preconditioners, while ensuring convergence by construction. As a preconditioner, we propose Unitary Neural Operators as a modification of Fourier Neural Operators. Our method can be interpreted as a data-driven discovery of Green's functions, which are then used to accelerate iterative solvers. We evaluate UNO-CG on various homogenization problems involving heterogeneous microstructures and millions of degrees of freedom. Our results demonstrate that UNO-CG enables a substantial reduction in the number of iterations and is competitive with handcrafted preconditioners for homogenization problems that involve expert knowledge. Moreover, UNO-CG maintains strong performance across a variety of boundary conditions, where many specialized solvers are not applicable, highlighting its versatility and robustness.
Authors: Chi-Sheng Chen, Aidan Hung-Wen Tsai
Abstract: The rise of decentralized finance (DeFi) has created a growing demand for accurate yield and performance forecasting to guide liquidity allocation strategies. In this study, we benchmark six models, XGBoost, Random Forest, LSTM, Transformer, quantum neural networks (QNN), and quantum support vector machines with quantum feature maps (QSVM-QNN), on one year of historical data from 28 Curve Finance pools. We evaluate model performance on test MAE, RMSE, and directional accuracy. Our results show that classical ensemble models, particularly XGBoost and Random Forest, consistently outperform both deep learning and quantum models. XGBoost achieves the highest directional accuracy (71.57%) with a test MAE of 1.80, while Random Forest attains the lowest test MAE of 1.77 and 71.36% accuracy. In contrast, quantum models underperform with directional accuracy below 50% and higher errors, highlighting current limitations in applying quantum machine learning to real-world DeFi time series data. This work offers a reproducible benchmark and practical insights into model suitability for DeFi applications, emphasizing the robustness of classical methods over emerging quantum approaches in this domain.
Authors: Jiawei Wang, Yu Guan, Chen Chen, Ligang Zhou, Laurence T. Yang, Sai Gu
Abstract: Sleep monitoring through accessible wearable technology is crucial to improving well-being in ubiquitous computing. Although photoplethysmography(PPG) sensors are widely adopted in consumer devices, achieving consistently reliable sleep staging using PPG alone remains a non-trivial challenge. In this work, we explore multiple strategies to enhance the performance of PPG-based sleep staging. Specifically, we compare conventional single-stream model with dual-stream cross-attention strategies, based on which complementary information can be learned via PPG and PPG-derived modalities such as augmented PPG or synthetic ECG. To study the effectiveness of the aforementioned approaches in four-stage sleep monitoring task, we conducted experiments on the world's largest sleep staging dataset, i.e., the Multi-Ethnic Study of Atherosclerosis(MESA). We found that substantial performance gain can be achieved by combining PPG and its auxiliary information under the dual-stream cross-attention architecture. Source code of this project can be found at https://github.com/DavyWJW/sleep-staging-models
Authors: Wenbo Cao, Weiwei Zhang
Abstract: Optimization-based PDE solvers that minimize scalar loss functions have gained increasing attention in recent years. These methods either define the loss directly over discrete variables, as in Optimizing a Discrete Loss (ODIL), or indirectly through a neural network surrogate, as in Physics-Informed Neural Networks (PINNs). However, despite their promise, such methods often converge much more slowly than classical iterative solvers and are commonly regarded as inefficient. This work provides a theoretical insight, attributing the inefficiency to the use of the mean squared error (MSE) loss, which implicitly forms the normal equations, squares the condition number, and severely impairs optimization. To address this, we propose a novel Stabilized Gradient Residual (SGR) loss. By tuning a weight parameter, it flexibly modulates the condition number between the original system and its normal equations, while reducing to the MSE loss in the limiting case. We systematically benchmark the convergence behavior and optimization stability of the SGR loss within both the ODIL framework and PINNs-employing either numerical or automatic differentiation-and compare its performance against classical iterative solvers. Numerical experiments on a range of benchmark problems demonstrate that, within the ODIL framework, the proposed SGR loss achieves orders-of-magnitude faster convergence than the MSE loss. Further validation within the PINNs framework shows that, despite the high nonlinearity of neural networks, SGR consistently outperforms the MSE loss. These theoretical and empirical findings help bridge the performance gap between classical iterative solvers and optimization-based solvers, highlighting the central role of loss conditioning, and provide key insights for the design of more efficient PDE solvers.
Authors: Ricardo Ribeiro Pereira, Jacopo Bono, Hugo Ferreira, Pedro Ribeiro, Carlos Soares, Pedro Bizarro
Abstract: When the available data for a target domain is limited, transfer learning (TL) methods can be used to develop models on related data-rich domains, before deploying them on the target domain. However, these TL methods are typically designed with specific, static assumptions on the amount of available labeled and unlabeled target data. This is in contrast with many real world applications, where the availability of data and corresponding labels varies over time. Since the evaluation of the TL methods is typically also performed under the same static data availability assumptions, this would lead to unrealistic expectations concerning their performance in real world settings. To support a more realistic evaluation and comparison of TL algorithms and models, we propose a data manipulation framework that (1) simulates varying data availability scenarios over time, (2) creates multiple domains through resampling of a given dataset and (3) introduces inter-domain variability by applying realistic domain transformations, e.g., creating a variety of potentially time-dependent covariate and concept shifts. These capabilities enable simulation of a large number of realistic variants of the experiments, in turn providing more information about the potential behavior of algorithms when deployed in dynamic settings. We demonstrate the usefulness of the proposed framework by performing a case study on a proprietary real-world suite of card payment datasets. Given the confidential nature of the case study, we also illustrate the use of the framework on the publicly available Bank Account Fraud (BAF) dataset. By providing a methodology for evaluating TL methods over time and in realistic data availability scenarios, our framework facilitates understanding of the behavior of models and algorithms. This leads to better decision making when deploying models for new domains in real-world environments.
Authors: Evangelos Sariyanidi, John D. Herrington, Lisa Yankowitz, Pratik Chaudhari, Theodore D. Satterthwaite, Casey J. Zampella, Robert T. Schultz, Russell T. Shinohara, Birkan Tunc
Abstract: Measuring the statistical dependence between observed signals is a primary tool for scientific discovery. However, biological systems often exhibit complex non-linear interactions that currently cannot be captured without a priori knowledge regarding the nature of dependence. We introduce a self-supervised approach, concurrence, which is inspired by the observation that if two signals are dependent, then one should be able to distinguish between temporally aligned vs. misaligned segments extracted from them. Experiments with fMRI, physiological and behavioral signals show that, to our knowledge, concurrence is the first approach that can expose relationships across such a wide spectrum of signals and extract scientifically relevant differences without ad-hoc parameter tuning or reliance on a priori information, providing a potent tool for scientific discoveries across fields. However, depencencies caused by extraneous factors remain an open problem, thus researchers should validate that exposed relationships truely pertain to the question(s) of interest.
Authors: Wei Li, Limei Hu, Feng Chen, Ye Yao
Abstract: In multi-task adversarial networks, the accurate estimation of unknown parameters in a distributed algorithm is hindered by attacked nodes or links. To tackle this challenge, this brief proposes a low-communication resilient distributed estimation algorithm. First, a node selection strategy based on reputation is introduced that allows nodes to communicate with more reliable subset of neighbors. Subsequently, to discern trustworthy intermediate estimates, the Weighted Support Vector Data Description (W-SVDD) model is employed to train the memory data. This trained model contributes to reinforce the resilience of the distributed estimation process against the impact of attacked nodes or links. Additionally, an event-triggered mechanism is introduced to minimize ineffective updates to the W-SVDD model, and a suitable threshold is derived based on assumptions. The convergence of the algorithm is analyzed. Finally, simulation results demonstrate that the proposed algorithm achieves superior performance with less communication cost compared to other algorithms.
Authors: Beatriz Macas Ord\'o\~nez, Diego Vinicio Orellana Villavicencio, Jos\'e Manuel Ferr\'andez, Paula Bonomini
Abstract: This study explores different neural network architectures to evaluate their ability to extract spatial and temporal patterns from electrocardiographic (ECG) signals and classify them into three groups: healthy subjects, Left Bundle Branch Block (LBBB), and Strict Left Bundle Branch Block (sLBBB). Clinical Relevance, Innovative technologies enable the selection of candidates for Cardiac Resynchronization Therapy (CRT) by optimizing the classification of subjects with Left Bundle Branch Block (LBBB).
Authors: Pallock Halder, Satyajit Mojumder
Abstract: Modern engineering systems are increasingly equipped with sensors for real-time monitoring and decision-making. However, the data collected by these sensors is often noisy and difficult to interpret, limiting its utility for control and diagnostics. In this work, we propose a physics-informed denoising framework that integrates energy-based model and Fisher score regularization to jointly reduce data noise and enforce physical consistency with a physics-based model. The approach is first validated on benchmark problems, including the simple harmonic oscillator, Burgers' equation, and Laplace's equation, across varying noise levels. We then apply the denoising framework to real thermal emission data from laser powder bed fusion (LPBF) additive manufacturing experiments, using a trained Physics-Informed Neural Network (PINN) surrogate model of the LPBF process to guide denoising. Results show that the proposed method outperforms baseline neural network denoisers, effectively reducing noise under a range of LPBF processing conditions. This physics-guided denoising strategy enables robust, real-time interpretation of low-cost sensor data, facilitating predictive control and improved defect mitigation in additive manufacturing.
Authors: Yahia Dalbah, Marcel Worring, Yen-Chia Hsu
Abstract: Urban air pollution is a major health crisis causing millions of premature deaths annually, underscoring the urgent need for accurate and scalable monitoring of air quality (AQ). While low-cost sensors (LCS) offer a scalable alternative to expensive reference-grade stations, their readings are affected by drift, calibration errors, and environmental interference. To address these challenges, we introduce Veli (Reference-free Variational Estimation via Latent Inference), an unsupervised Bayesian model that leverages variational inference to correct LCS readings without requiring co-location with reference stations, eliminating a major deployment barrier. Specifically, Veli constructs a disentangled representation of the LCS readings, effectively separating the true pollutant reading from the sensor noise. To build our model and address the lack of standardized benchmarks in AQ monitoring, we also introduce the Air Quality Sensor Data Repository (AQ-SDR). AQ-SDR is the largest AQ sensor benchmark to date, with readings from 23,737 LCS and reference stations across multiple regions. Veli demonstrates strong generalization across both in-distribution and out-of-distribution settings, effectively handling sensor drift and erratic sensor behavior. Code for model and dataset will be made public when this paper is published.
Authors: Lucio Pinello, Francesco Cadini, Luca Lomazzi
Abstract: Ultrasonic Guided Waves (UGWs) represent a promising diagnostic tool for Structural Health Monitoring (SHM) in thin-walled structures, and their integration with machine learning (ML) algorithms is increasingly being adopted to enable real-time monitoring capabilities. However, the large-scale deployment of UGW-based ML methods is constrained by data scarcity and limited generalisation across different materials and sensor configurations. To address these limitations, this work proposes a novel transfer learning (TL) framework based on Multilinear Principal Component Analysis (MPCA). First, a Convolutional Neural Network (CNN) for regression is trained to perform damage localisation for a plated structure. Then, MPCA and fine-tuning are combined to have the CNN work for a different plate. By jointly applying MPCA to the source and target domains, the method extracts shared latent features, enabling effective domain adaptation without requiring prior assumptions about dimensionality. Following MPCA, fine-tuning enables adapting the pre-trained CNN to a new domain without the need for a large training dataset. The proposed MPCA-based TL method was tested against 12 case studies involving different composite materials and sensor arrays. Statistical metrics were used to assess domains alignment both before and after MPCA, and the results demonstrate a substantial reduction in localisation error compared to standard TL techniques. Hence, the proposed approach emerges as a robust, data-efficient, and statistically based TL framework for UGW-based SHM.
Authors: Yumeng Shi, Zhongliang Yang, DiYang Lu, Yisi Wang, Yiting Zhou, Linna Zhou
Abstract: Corporate credit rating serves as a crucial intermediary service in the market economy, playing a key role in maintaining economic order. Existing credit rating models rely on financial metrics and deep learning. However, they often overlook insights from non-financial data, such as corporate annual reports. To address this, this paper introduces a corporate credit rating framework that integrates financial data with features extracted from annual reports using FinBERT, aiming to fully leverage the potential value of unstructured text data. In addition, we have developed a large-scale dataset, the Comprehensive Corporate Rating Dataset (CCRD), which combines both traditional financial data and textual data from annual reports. The experimental results show that the proposed method improves the accuracy of the rating predictions by 8-12%, significantly improving the effectiveness and reliability of corporate credit ratings.
Authors: Yu Shi, Zongliang Fu, Shuo Chen, Bohan Zhao, Wei Xu, Changshui Zhang, Jian Li
Abstract: The success of large-scale pre-training paradigm, exemplified by Large Language Models (LLMs), has inspired the development of Time Series Foundation Models (TSFMs). However, their application to financial candlestick (K-line) data remains limited, often underperforming non-pre-trained architectures. Moreover, existing TSFMs often overlook crucial downstream tasks such as volatility prediction and synthetic data generation. To address these limitations, we propose Kronos, a unified, scalable pre-training framework tailored to financial K-line modeling. Kronos introduces a specialized tokenizer that discretizes continuous market information into token sequences, preserving both price dynamics and trade activity patterns. We pre-train Kronos using an autoregressive objective on a massive, multi-market corpus of over 12 billion K-line records from 45 global exchanges, enabling it to learn nuanced temporal and cross-asset representations. Kronos excels in a zero-shot setting across a diverse set of financial tasks. On benchmark datasets, Kronos boosts price series forecasting RankIC by 93% over the leading TSFM and 87% over the best non-pre-trained baseline. It also achieves a 9% lower MAE in volatility forecasting and a 22% improvement in generative fidelity for synthetic K-line sequences. These results establish Kronos as a robust, versatile foundation model for end-to-end financial time series analysis. Our pre-trained model is publicly available at https://github.com/shiyu-coder/Kronos.
Authors: Chunyu Liu, Hao Zhang, Wei Wu, Fuhui Zhou, Qihui Wu, Derrick Wing Kwan Ng, Chan-Byoung Chae
Abstract: The enhancement of spectrum efficiency and the realization of secure spectrum utilization are critically dependent on spectrum cognition. However, existing spectrum cognition methods often exhibit limited generalization and suboptimal accuracy when deployed across diverse spectrum environments and tasks. To overcome these challenges, we propose a spectrum foundation model, termed SpectrumFM, which provides a new paradigm for spectrum cognition. An innovative spectrum encoder that exploits the convolutional neural networks and the multi-head self attention mechanisms is proposed to effectively capture both fine-grained local signal structures and high-level global dependencies in the spectrum data. To enhance its adaptability, two novel self-supervised learning tasks, namely masked reconstruction and next-slot signal prediction, are developed for pre-training SpectrumFM, enabling the model to learn rich and transferable representations. Furthermore, low-rank adaptation (LoRA) parameter-efficient fine-tuning is exploited to enable SpectrumFM to seamlessly adapt to various downstream spectrum cognition tasks, including spectrum sensing (SS), anomaly detection (AD), and wireless technology classification (WTC). Extensive experiments demonstrate the superiority of SpectrumFM over state-of-the-art methods. Specifically, it improves detection probability in the SS task by 30% at -4 dB signal-to-noise ratio (SNR), boosts the area under the curve (AUC) in the AD task by over 10%, and enhances WTC accuracy by 9.6%.
Authors: Vinil Polepalli
Abstract: Pan-cancer classification using transcriptomic (RNA-Seq) data can inform tumor subtyping and therapy selection, but is challenging due to extremely high dimensionality and limited sample sizes. In this study, we propose a novel deep learning framework that uses a class-conditional variational autoencoder (cVAE) to augment training data for pan-cancer gene expression classification. Using 801 tumor RNA-Seq samples spanning 5 cancer types from The Cancer Genome Atlas (TCGA), we first perform feature selection to reduce 20,531 gene expression features to the 500 most variably expressed genes. A cVAE is then trained on this data to learn a latent representation of gene expression conditioned on cancer type, enabling the generation of synthetic gene expression samples for each tumor class. We augment the training set with these cVAE-generated samples (doubling the dataset size) to mitigate overfitting and class imbalance. A two-layer multilayer perceptron (MLP) classifier is subsequently trained on the augmented dataset to predict tumor type. The augmented framework achieves high classification accuracy (~98%) on a held-out test set, substantially outperforming a classifier trained on the original data alone. We present detailed experimental results, including VAE training curves, classifier performance metrics (ROC curves and confusion matrix), and architecture diagrams to illustrate the approach. The results demonstrate that cVAE-based synthetic augmentation can significantly improve pan-cancer prediction performance, especially for underrepresented cancer classes.
Authors: Yihao Ang, Qiang Wang, Qiang Huang, Yifan Bao, Xinyu Xi, Anthony K. H. Tung, Chen Jin, Zhiyong Huang
Abstract: Synthetic time series are essential tools for data augmentation, stress testing, and algorithmic prototyping in quantitative finance. However, in cryptocurrency markets, characterized by 24/7 trading, extreme volatility, and rapid regime shifts, existing Time Series Generation (TSG) methods and benchmarks often fall short, jeopardizing practical utility. Most prior work (1) targets non-financial or traditional financial domains, (2) focuses narrowly on classification and forecasting while neglecting crypto-specific complexities, and (3) lacks critical financial evaluations, particularly for trading applications. To address these gaps, we introduce \textsf{CTBench}, the first comprehensive TSG benchmark tailored for the cryptocurrency domain. \textsf{CTBench} curates an open-source dataset from 452 tokens and evaluates TSG models across 13 metrics spanning 5 key dimensions: forecasting accuracy, rank fidelity, trading performance, risk assessment, and computational efficiency. A key innovation is a dual-task evaluation framework: (1) the \emph{Predictive Utility} task measures how well synthetic data preserves temporal and cross-sectional patterns for forecasting, while (2) the \emph{Statistical Arbitrage} task assesses whether reconstructed series support mean-reverting signals for trading. We benchmark eight representative models from five methodological families over four distinct market regimes, uncovering trade-offs between statistical fidelity and real-world profitability. Notably, \textsf{CTBench} offers model ranking analysis and actionable guidance for selecting and deploying TSG models in crypto analytics and strategy development.
Authors: Eduardo Abi Jaber, Louis-Amand G\'erard
Abstract: We investigate the use of path signatures in a machine learning context for hedging exotic derivatives under non-Markovian stochastic volatility models. In a deep learning setting, we use signatures as features in feedforward neural networks and show that they outperform LSTMs in most cases, with orders of magnitude less training compute. In a shallow learning setting, we compare two regression approaches: the first directly learns the hedging strategy from the expected signature of the price process; the second models the dynamics of volatility using a signature volatility model, calibrated on the expected signature of the volatility. Solving the hedging problem in the calibrated signature volatility model yields more accurate and stable results across different payoffs and volatility dynamics.
Authors: Kirill Sokolov, Alexander Korotin
Abstract: We consider the discrete-time Schr\"odinger bridge problem on a finite state space. Although it has been known that the Iterative Markovian Fitting (IMF) algorithm converges in Kullback-Leibler divergence to the ground truth solution, the speed of that convergence remained unquantified. In this work, we establish for the first time that IMF exhibits exponential convergence with an explicit contraction factor.
Authors: Zongyou Yang, Jonathan Loo
Abstract: Recently, a significant improvement in the accuracy of 3D human pose estimation has been achieved by combining convolutional neural networks (CNNs) with pyramid grid alignment feedback loops. Additionally, innovative breakthroughs have been made in the field of computer vision through the adoption of Transformer-based temporal analysis architectures. Given these advancements, this study aims to deeply optimize and improve the existing Pymaf network architecture. The main innovations of this paper include: (1) Introducing a Transformer feature extraction network layer based on self-attention mechanisms to enhance the capture of low-level features; (2) Enhancing the understanding and capture of temporal signals in video sequences through feature temporal fusion techniques; (3) Implementing spatial pyramid structures to achieve multi-scale feature fusion, effectively balancing feature representations differences across different scales. The new PyCAT4 model obtained in this study is validated through experiments on the COCO and 3DPW datasets. The results demonstrate that the proposed improvement strategies significantly enhance the network's detection capability in human pose estimation, further advancing the development of human pose estimation technology.
Authors: Radhika Dua (Fred), Young Joon (Fred), Kwon, Siddhant Dogra, Daniel Freedman, Diana Ruan, Motaz Nashawaty, Danielle Rigau, Daniel Alexander Alber, Kang Zhang, Kyunghyun Cho, Eric Karl Oermann
Abstract: Radiological imaging is central to diagnosis, treatment planning, and clinical decision-making. Vision-language foundation models have spurred interest in automated radiology report generation (RRG), but safe deployment requires reliable clinical evaluation of generated reports. Existing metrics often rely on surface-level similarity or behave as black boxes, lacking interpretability. We introduce ICARE (Interpretable and Clinically-grounded Agent-based Report Evaluation), an interpretable evaluation framework leveraging large language model agents and dynamic multiple-choice question answering (MCQA). Two agents, each with either the ground-truth or generated report, generate clinically meaningful questions and quiz each other. Agreement on answers captures preservation and consistency of findings, serving as interpretable proxies for clinical precision and recall. By linking scores to question-answer pairs, ICARE enables transparent, and interpretable assessment. Clinician studies show ICARE aligns significantly more with expert judgment than prior metrics. Perturbation analyses confirm sensitivity to clinical content and reproducibility, while model comparisons reveal interpretable error patterns.
Authors: Mengyu Zhang, Zhuotao Liu, Jingwen Huang, Xuanqi Liu
Abstract: Privacy-preserving machine learning (PPML) is critical to ensure data privacy in AI. Over the past few years, the community has proposed a wide range of provably secure PPML schemes that rely on various cryptography primitives. However, when it comes to large language models (LLMs) with billions of parameters, the efficiency of PPML is everything but acceptable. For instance, the state-of-the-art solution for confidential LLM inference represents at least 10,000-fold slower performance compared to plaintext inference. The performance gap is even larger when the context length increases. In this position paper, we propose a novel framework named Agentic-PPML to make PPML in LLMs practical. Our key insight is to employ a general-purpose LLM for intent understanding and delegate cryptographically secure inference to specialized models trained on vertical domains. By modularly separating language intent parsing - which typically involves little or no sensitive information - from privacy-critical computation, Agentic-PPML completely eliminates the need for the LLMs to process the encrypted prompts, enabling practical deployment of privacy-preserving LLM-centric services.
Authors: J. Alex Hurt, Trevor M. Bajkowski, Grant J. Scott, Curt H. Davis
Abstract: In 2012, AlexNet established deep convolutional neural networks (DCNNs) as the state-of-the-art in CV, as these networks soon led in visual tasks for many domains, including remote sensing. With the publication of Visual Transformers, we are witnessing the second modern leap in computational vision, and as such, it is imperative to understand how various transformer-based neural networks perform on satellite imagery. While transformers have shown high levels of performance in natural language processing and CV applications, they have yet to be compared on a large scale to modern remote sensing data. In this paper, we explore the use of transformer-based neural networks for object detection in high-resolution electro-optical satellite imagery, demonstrating state-of-the-art performance on a variety of publicly available benchmark data sets. We compare eleven distinct bounding-box detection and localization algorithms in this study, of which seven were published since 2020, and all eleven since 2015. The performance of five transformer-based architectures is compared with six convolutional networks on three state-of-the-art opensource high-resolution remote sensing imagery datasets ranging in size and complexity. Following the training and evaluation of thirty-three deep neural models, we then discuss and analyze model performance across various feature extraction methodologies and detection algorithms.
Authors: Giovanni Cherubin, Andrew Paverd
Abstract: Preventing jailbreaking and model hijacking of Large Language Models (LLMs) is an important yet challenging task. For example, when interacting with a chatbot, malicious users can input specially crafted prompts to cause the LLM to generate undesirable content or perform a completely different task from its intended purpose. Existing mitigations for such attacks typically rely on hardening the LLM's system prompt or using a content classifier trained to detect undesirable content or off-topic conversations. However, these probabilistic approaches are relatively easy to bypass due to the very large space of possible inputs and undesirable outputs. In this paper, we present and evaluate Highlight & Summarize (H&S), a new design pattern for retrieval-augmented generation (RAG) systems that prevents these attacks by design. The core idea is to perform the same task as a standard RAG pipeline (i.e., to provide natural language answers to questions, based on relevant sources) without ever revealing the user's question to the generative LLM. This is achieved by splitting the pipeline into two components: a highlighter, which takes the user's question and extracts relevant passages ("highlights") from the retrieved documents, and a summarizer, which takes the highlighted passages and summarizes them into a cohesive answer. We describe several possible instantiations of H&S and evaluate their generated responses in terms of correctness, relevance, and response quality. Surprisingly, when using an LLM-based highlighter, the majority of H&S responses are judged to be better than those of a standard RAG pipeline.
Authors: Brennen A. Hill, Mant Koh En Wei, Thangavel Jishnuanandh
Abstract: We compare the efficacy of learned versus engineered communication strategies in a cooperative multi-agent reinforcement learning (MARL) environment. For the learned approach, we introduce Learned Direct Communication (LDC), where agents generate messages and actions concurrently via a neural network. Our engineered approach, Intention Communication, employs an Imagined Trajectory Generation Module (ITGM) and a Message Generation Network (MGN) to formulate messages based on predicted future states. Both strategies are evaluated on their success rates in cooperative tasks under fully and partially observable conditions. Our findings indicate that while emergent communication is viable, the engineered approach demonstrates superior performance and scalability, particularly as environmental complexity increases.
Authors: Dai Li, Kevin Course, Wei Li, Hongwei Li, Jie Hua, Yiqi Chen, Zhao Zhu, Rui Jian, Xuan Cao, Bi Xue, Yu Shi, Jing Qian, Kai Ren, Matt Ma, Qunshu Zhang, Rui Li
Abstract: While scaling laws promise significant performance gains for recommender systems, efficiently deploying hyperscale models remains a major unsolved challenge. In contrast to fields where FMs are already widely adopted such as natural language processing and computer vision, progress in recommender systems is hindered by unique challenges including the need to learn from online streaming data under shifting data distributions, the need to adapt to different recommendation surfaces with a wide diversity in their downstream tasks and their input distributions, and stringent latency and computational constraints. To bridge this gap, we propose to leverage the Foundation-Expert Paradigm: a framework designed for the development and deployment of hyperscale recommendation FMs. In our approach, a central FM is trained on lifelong, cross-surface, multi-modal user data to learn generalizable knowledge. This knowledge is then efficiently transferred to various lightweight, surface-specific ``expert" models via target-aware embeddings, allowing them to adapt to local data distributions and optimization goals with minimal overhead. To meet our training, inference and development needs, we built HyperCast, a production-grade infrastructure system that re-engineers training, serving, logging and iteration to power this decoupled paradigm. Our approach is now deployed at Meta serving tens of billions of user requests daily, demonstrating online metric improvements over our previous one-stage production system while improving developer velocity and maintaining infrastructure efficiency. To the best of our knowledge, this work represents the first successful deployment of a Foundation-Expert paradigm at this scale, offering a proven, compute-efficient, and developer-friendly blueprint to realize the promise of scaling laws in recommender systems.
Authors: Ilias Aarab
Abstract: Bank supervisors face the complex task of ensuring that new measures are consistently aligned with historical precedents. To address this challenge, we introduce a novel Information Retrieval (IR) System tailored to assist supervisors in drafting both consistent and effective measures. This system ingests findings from on-site investigations. It then retrieves the most relevant historical findings and their associated measures from a comprehensive database, providing a solid basis for supervisors to write well-informed measures for new findings. Utilizing a blend of lexical, semantic, and Capital Requirements Regulation (CRR) fuzzy set matching techniques, the IR system ensures the retrieval of findings that closely align with current cases. The performance of this system, particularly in scenarios with partially labeled data, is validated through a Monte Carlo methodology, showcasing its robustness and accuracy. Enhanced by a Transformer-based Denoising AutoEncoder for fine-tuning, the final model achieves a Mean Average Precision (MAP@100) of 0.83 and a Mean Reciprocal Rank (MRR@100) of 0.92. These scores surpass those of both standalone lexical models such as BM25 and semantic BERT-like models.
Authors: Alireza Ghafarollahi, Markus J. Buehler
Abstract: Conventional machine learning approaches accelerate inorganic materials design via accurate property prediction and targeted material generation, yet they operate as single-shot models limited by the latent knowledge baked into their training data. A central challenge lies in creating an intelligent system capable of autonomously executing the full inorganic materials discovery cycle, from ideation and planning to experimentation and iterative refinement. We introduce SparksMatter, a multi-agent AI model for automated inorganic materials design that addresses user queries by generating ideas, designing and executing experimental workflows, continuously evaluating and refining results, and ultimately proposing candidate materials that meet the target objectives. SparksMatter also critiques and improves its own responses, identifies research gaps and limitations, and suggests rigorous follow-up validation steps, including DFT calculations and experimental synthesis and characterization, embedded in a well-structured final report. The model's performance is evaluated across case studies in thermoelectrics, semiconductors, and perovskite oxides materials design. The results demonstrate the capacity of SparksMatter to generate novel stable inorganic structures that target the user's needs. Benchmarking against frontier models reveals that SparksMatter consistently achieves higher scores in relevance, novelty, and scientific rigor, with a significant improvement in novelty across multiple real-world design tasks as assessed by a blinded evaluator. These results demonstrate SparksMatter's unique capacity to generate chemically valid, physically meaningful, and creative inorganic materials hypotheses beyond existing materials knowledge.
Authors: Chia-Tung Ho, Jing Gong, Xufeng Yao, Yunsheng Bai, Abhishek B Akkur, Haoxing Ren
Abstract: Large language models (LLMs) excel at solving complex tasks by executing agentic workflows composed of detailed instructions and structured operations. Yet, building general-purpose agents by manually embedding foundation models into agentic systems such as Chain-of-Thought, Self-Reflection, and ReACT through text interfaces limits scalability and efficiency. Recently, many researchers have sought to automate the generation and optimization of these workflows through code-based representations. However, existing methods often rely on labeled datasets to train and optimize workflows, making them ineffective and inflexible for solving real-world, dynamic problems where labeled data is unavailable. To address this challenge, we introduce Polymath, a self-optimizing agent with dynamic hierarchical workflow that leverages the flexibility of task flow graphs and the expressiveness of code-represented workflows to solve a wide range of real-world, dynamic problems. The proposed optimization methodology integrates multi-grid-inspired graph optimization with a self-reflection-guided evolutionary algorithm to refine workflows without labeled data. Experimental results on six benchmark datasets across coding, math, and multi-turn QA tasks show that Polymath achieves 8.1% average improvement over state-of-the-art baselines.
Authors: Peng Ding, Rick Stevens
Abstract: The proliferation of tool-augmented Large Language Models (LLMs) has created a fragmented ecosystem where developers must navigate multiple protocols, manual schema definitions, and complex execution workflows. We address this challenge by proposing a unified approach to tool integration that abstracts protocol differences while optimizing execution performance. Our solution demonstrates how protocol-agnostic design principles can significantly reduce development overhead through automated schema generation, dual-mode concurrent execution, and seamless multi-source tool management. Experimental results show 60-80% code reduction across integration scenarios, performance improvements up to 3.1x through optimized concurrency, and full compatibility with existing function calling standards. This work contributes both theoretical insights into tool integration architecture and practical solutions for real-world LLM application development.
Authors: Brennen A. Hill, Zhang Xinyu, Timothy Putra Prasetio
Abstract: Despite their success in image classification, modern convolutional neural networks (CNNs) exhibit fundamental limitations, including data inefficiency, poor out-of-distribution generalization, and vulnerability to adversarial perturbations. The primate visual system, in contrast, demonstrates superior efficiency and robustness, suggesting that its architectural principles may offer a blueprint for more capable artificial vision systems. This paper introduces Visual Cortex Network (VCNet), a novel neural network architecture whose design is informed by the macro-scale organization of the primate visual cortex. VCNet emulates key biological mechanisms, including hierarchical processing across distinct cortical areas, dual-stream information segregation, and top-down predictive feedback. We evaluate VCNet on two specialized benchmarks: the Spots-10 animal pattern dataset and a light field image classification task. Our results show that VCNet achieves a classification accuracy of 92.1\% on Spots-10 and 74.4\% on the light field dataset, surpassing contemporary models of comparable size. This work demonstrates that integrating neuroscientific principles into network design can lead to more efficient and robust models, providing a promising direction for addressing long-standing challenges in machine learning.
Authors: Malek Itani, Ashton Graves, Sefik Emre Eskimez, Shyamnath Gollakota
Abstract: We present the first neural target speech extraction (TSE) system that uses human feedback for iterative refinement. Our approach allows users to mark specific segments of the TSE output, generating an edit mask. The refinement system then improves the marked sections while preserving unmarked regions. Since large-scale datasets of human-marked errors are difficult to collect, we generate synthetic datasets using various automated masking functions and train models on each. Evaluations show that models trained with noise power-based masking (in dBFS) and probabilistic thresholding perform best, aligning with human annotations. In a study with 22 participants, users showed a preference for refined outputs over baseline TSE. Our findings demonstrate that human-in-the-loop refinement is a promising approach for improving the performance of neural speech extraction.
Authors: Yi-Hsuan Hsiao, Andrea Tagliabue, Owen Matteson, Suhan Kim, Tong Zhao, Jonathan P. How, YuFeng Chen
Abstract: Aerial insects exhibit highly agile maneuvers such as sharp braking, saccades, and body flips under disturbance. In contrast, insect-scale aerial robots are limited to tracking non-aggressive trajectories with small body acceleration. This performance gap is contributed by a combination of low robot inertia, fast dynamics, uncertainty in flapping-wing aerodynamics, and high susceptibility to environmental disturbance. Executing highly dynamic maneuvers requires the generation of aggressive flight trajectories that push against the hardware limit and a high-rate feedback controller that accounts for model and environmental uncertainty. Here, through designing a deep-learned robust tube model predictive controller, we showcase insect-like flight agility and robustness in a 750-millgram flapping-wing robot. Our model predictive controller can track aggressive flight trajectories under disturbance. To achieve a high feedback rate in a compute-constrained real-time system, we design imitation learning methods to train a two-layer, fully connected neural network, which resembles insect flight control architecture consisting of central nervous system and motor neurons. Our robot demonstrates insect-like saccade movements with lateral speed and acceleration of 197 centimeters per second and 11.7 meters per second square, representing 447$\%$ and 255$\%$ improvement over prior results. The robot can also perform saccade maneuvers under 160 centimeters per second wind disturbance and large command-to-force mapping errors. Furthermore, it performs 10 consecutive body flips in 11 seconds - the most challenging maneuver among sub-gram flyers. These results represent a milestone in achieving insect-scale flight agility and inspire future investigations on sensing and compute autonomy.
Authors: Malek Itani, Tuochao Chen, Shyamnath Gollakota
Abstract: Speech separation on hearable devices can enable transformative augmented and enhanced hearing capabilities. However, state-of-the-art speech separation networks cannot run in real-time on tiny, low-power neural accelerators designed for hearables, due to their limited compute capabilities. We present TF-MLPNet, the first speech separation network capable of running in real-time on such low-power accelerators while outperforming existing streaming models for blind speech separation and target speech extraction. Our network operates in the time-frequency domain, processing frequency sequences with stacks of fully connected layers that alternate along the channel and frequency dimensions, and independently processing the time sequence at each frequency bin using convolutional layers. Results show that our mixed-precision quantization-aware trained (QAT) model can process 6 ms audio chunks in real-time on the GAP9 processor, achieving a 3.5-4x runtime reduction compared to prior speech separation models.
Authors: Anuroop Sriram, Logan M. Brabson, Xiaohan Yu, Sihoon Choi, Kareem Abdelmaqsoud, Elias Moubarak, Pim de Haan, Sindy L\"owe, Johann Brehmer, John R. Kitchin, Max Welling, C. Lawrence Zitnick, Zachary Ulissi, Andrew J. Medford, David S. Sholl
Abstract: Identifying useful sorbent materials for direct air capture (DAC) from humid air remains a challenge. We present the Open DAC 2025 (ODAC25) dataset, a significant expansion and improvement upon ODAC23 (Sriram et al., ACS Central Science, 10 (2024) 923), comprising nearly 70 million DFT single-point calculations for CO$_2$, H$_2$O, N$_2$, and O$_2$ adsorption in 15,000 MOFs. ODAC25 introduces chemical and configurational diversity through functionalized MOFs, high-energy GCMC-derived placements, and synthetically generated frameworks. ODAC25 also significantly improves upon the accuracy of DFT calculations and the treatment of flexible MOFs in ODAC23. Along with the dataset, we release new state-of-the-art machine-learned interatomic potentials trained on ODAC25 and evaluate them on adsorption energy and Henry's law coefficient predictions.
Authors: Mohammed Salah Al-Radhi, G\'eza N\'emeth, Branislav Gerazov
Abstract: Speech synthesis from intracranial EEG (iEEG) signals offers a promising avenue for restoring communication in individuals with severe speech impairments. However, achieving intelligible and natural speech remains challenging due to limitations in feature representation, prosody modeling, and phase reconstruction. We introduce MiSTR, a deep-learning framework that integrates: 1) Wavelet-based feature extraction to capture fine-grained temporal, spectral, and neurophysiological representations of iEEG signals, 2) A Transformer-based decoder for prosody-aware spectrogram prediction, and 3) A neural phase vocoder enforcing harmonic consistency via adaptive spectral correction. Evaluated on a public iEEG dataset, MiSTR achieves state-of-the-art speech intelligibility, with a mean Pearson correlation of 0.91 between reconstructed and original Mel spectrograms, improving over existing neural speech synthesis baselines.
Authors: Chenyang Wang, Liang Wen, Shousheng Jia, Xiangzheng Zhang, Liang Xu
Abstract: While advancements in the reasoning abilities of LLMs have significantly enhanced their performance in solving mathematical problems, coding tasks, and general puzzles, their effectiveness in accurately adhering to instructions remains inconsistent, particularly with more complex directives. Our investigation identifies lazy reasoning during the thinking stage as the primary factor contributing to poor instruction adherence. To mitigate this issue, we propose a comprehensive framework designed to enable rigorous reasoning processes involving preview and self-checking, essential for satisfying strict instruction constraints. Specifically, we first generate instructions with complex constraints and apply a filtering process to obtain valid prompts, resulting in three distinct prompt datasets categorized as hard, easy, and pass. Then, we employ rejection sampling on the pass prompts to curate a small yet high-quality dataset, enabling a cold-start initialization of the model and facilitating its adaptation to effective reasoning patterns. Subsequently, we employ an entropy-preserving supervised fine-tuning (Entropy-SFT) strategy coupled with token-wise entropy-adaptive (TEA-RL) reinforcement learning guided by rule-based dense rewards. This approach encourages the model to transform its reasoning mechanism, ultimately fostering generalizable reasoning abilities that encompass preview and self-checking. Extensive experiments conducted on instruction-following benchmarks demonstrate remarkable performance improvements across various model scales. Notably, our Light-IF-32B model surpasses both larger open-source models such as DeepSeek-R1 and closed-source models like Doubao-1.6.
Authors: Lukas P\"atz, Moritz Beyer, Jannik Sp\"ath, Lasse Bohlen, Patrick Zschech, Mathias Kraus, Julian Rosenberger
Abstract: This study investigates political discourse in the German parliament, the Bundestag, by analyzing approximately 28,000 parliamentary speeches from the last five years. Two machine learning models for topic and sentiment classification were developed and trained on a manually labeled dataset. The models showed strong classification performance, achieving an area under the receiver operating characteristic curve (AUROC) of 0.94 for topic classification (average across topics) and 0.89 for sentiment classification. Both models were applied to assess topic trends and sentiment distributions across political parties and over time. The analysis reveals remarkable relationships between parties and their role in parliament. In particular, a change in style can be observed for parties moving from government to opposition. While ideological positions matter, governing responsibilities also shape discourse. The analysis directly addresses key questions about the evolution of topics, sentiment dynamics, and party-specific discourse strategies in the Bundestag.
Authors: Bronya Roni Chernyak, Yael Segal, Yosi Shrem, Joseph Keshet
Abstract: Deep learning models excel at many tasks but rely on the assumption that training and test data follow the same distribution. This assumption often does not hold in real-world speech systems, where distribution shifts are common due to varying environments, recording conditions, and speaker diversity. The method of Domain Shifts with Uncertainty (DSU) augments the input of each neural network layer based on the input feature statistics. It addresses the problem of out-of-domain generalization by assuming feature statistics follow a multivariate Gaussian distribution and substitutes the input with sampled features from this distribution. While effective for computer vision, applying DSU to speech presents challenges due to the nature of the data. Unlike static visual data, speech is a temporal signal commonly represented by a spectrogram - the change of frequency over time. This representation cannot be treated as a simple image, and the resulting sparsity can lead to skewed feature statistics when applied to the entire input. To tackle out-of-distribution issues in keyword spotting, we propose PatchDSU, which extends DSU by splitting the input into patches and independently augmenting each patch. We evaluated PatchDSU and DSU alongside other methods on the Google Speech Commands, Librispeech, and TED-LIUM. Additionally, we evaluated performance under white Gaussian and MUSAN music noise conditions. We also explored out-of-domain generalization by analyzing model performance on datasets they were not trained on. Overall, in most cases, both PatchDSU and DSU outperform other methods. Notably, PatchDSU demonstrates more consistent improvements across the evaluated scenarios compared to other approaches.
Authors: Chuanzhi Xu, Haoxian Zhou, Langyi Chen, Yuk Ying Chung, Qiang Qu
Abstract: Event cameras offer unparalleled advantages such as high temporal resolution, low latency, and high dynamic range. However, their limited spatial resolution poses challenges for fine-grained perception tasks. In this work, we propose an ultra-lightweight, stream-based event-to-event super-resolution method based on Spiking Neural Networks (SNNs), designed for real-time deployment on resource-constrained devices. To further reduce model size, we introduce a novel Dual-Forward Polarity-Split Event Encoding strategy that decouples positive and negative events into separate forward paths through a shared SNN. Furthermore, we propose a Learnable Spatio-temporal Polarity-aware Loss (LearnSTPLoss) that adaptively balances temporal, spatial, and polarity consistency using learnable uncertainty-based weights. Experimental results demonstrate that our method achieves competitive super-resolution performance on multiple datasets while significantly reducing model size and inference time. The lightweight design enables embedding the module into event cameras or using it as an efficient front-end preprocessing for downstream vision tasks.
Authors: Anqi Li, Wenwei Jin, Jintao Tong, Pengda Qin, Weijia Li, Guo Lu
Abstract: Social platforms have revolutionized information sharing, but also accelerated the dissemination of harmful and policy-violating content. To ensure safety and compliance at scale, moderation systems must go beyond efficiency and offer accuracy and interpretability. However, current approaches largely rely on noisy, label-driven learning, lacking alignment with moderation rules and producing opaque decisions that hinder human review. Therefore, we propose Hierarchical Guard (Hi-Guard), a multimodal moderation framework that introduces a new policy-aligned decision paradigm. The term "Hierarchical" reflects two key aspects of our system design: (1) a hierarchical moderation pipeline, where a lightweight binary model first filters safe content and a stronger model handles fine-grained risk classification; and (2) a hierarchical taxonomy in the second stage, where the model performs path-based classification over a hierarchical taxonomy ranging from coarse to fine-grained levels. To ensure alignment with evolving moderation policies, Hi-Guard directly incorporates rule definitions into the model prompt. To further enhance structured prediction and reasoning, we introduce a multi-level soft-margin reward and optimize with Group Relative Policy Optimization (GRPO), penalizing semantically adjacent misclassifications and improving explanation quality. Extensive experiments and real-world deployment demonstrate that Hi-Guard achieves superior classification accuracy, generalization, and interpretability, paving the way toward scalable, transparent, and trustworthy content safety systems. Code is available at: https://github.com/lianqi1008/Hi-Guard.
Authors: Francisco Daunas, I\~naki Esnaola, Samir M. Perlaza
Abstract: The dual formulation of empirical risk minimization with f-divergence regularization (ERM-fDR) is introduced. The solution of the dual optimization problem to the ERM-fDR is connected to the notion of normalization function introduced as an implicit function. This dual approach leverages the Legendre-Fenchel transform and the implicit function theorem to provide a nonlinear ODE expression to the normalization function. Furthermore, the nonlinear ODE expression and its properties provide a computationally efficient method to calculate the normalization function of the ERM-fDR solution under a mild condition.
Authors: Matteo Caligiuri, Francesco Barbato, Donald Shenaj, Umberto Michieli, Pietro Zanuttigh
Abstract: Federated Learning (FL) is an established paradigm for training deep learning models on decentralized data. However, as the size of the models grows, conventional FL approaches often require significant computational resources on client devices, which may not be feasible. We introduce FedPromo, a novel framework that enables efficient adaptation of large-scale foundation models stored on a central server to new domains encountered only by remote clients. Instead of directly training the large model on client devices, FedPromo optimizes lightweight proxy models via FL, significantly reducing computational overhead while maintaining privacy. Our method follows a two-stage process: first, server-side knowledge distillation aligns the representations of a large-scale foundation model (e.g., a transformer) with those of a compact counterpart (e.g., a CNN). Then, the compact model encoder is deployed to client devices, where trainable classifiers are learned locally. These classifiers are subsequently aggregated and seamlessly transferred back to the foundation model, facilitating personalized adaptation without requiring direct access to user data. Through novel regularization strategies, our framework enables decentralized multi-domain learning, balancing performance, privacy, and resource efficiency. Extensive experiments on five image classification benchmarks demonstrate that FedPromo outperforms existing methods while assuming limited-resource clients.
Authors: Michael K. Chen
Abstract: General logical reasoning, defined as the ability to reason deductively on domain-agnostic tasks, continues to be a challenge for large language models (LLMs). Current LLMs fail to reason deterministically and are not interpretable. As such, there has been a recent surge in interest in neurosymbolic AI, which attempts to incorporate logic into neural networks. We first identify two main neurosymbolic approaches to improving logical reasoning: (i) the integrative approach comprising models where symbolic reasoning is contained within the neural network, and (ii) the hybrid approach comprising models where a symbolic solver, separate from the neural network, performs symbolic reasoning. Both contain AI systems with promising results on domain-specific logical reasoning benchmarks. However, their performance on domain-agnostic benchmarks is understudied. To the best of our knowledge, there has not been a comparison of the contrasting approaches that answers the following question: Which approach is more promising for developing general logical reasoning? To analyze their potential, the following best-in-class domain-agnostic models are introduced: Logic Neural Network (LNN), which uses the integrative approach, and LLM-Symbolic Solver (LLM-SS), which uses the hybrid approach. Using both models as case studies and representatives of each approach, our analysis demonstrates that the hybrid approach is more promising for developing general logical reasoning because (i) its reasoning chain is more interpretable, and (ii) it retains the capabilities and advantages of existing LLMs. To support future works using the hybrid approach, we propose a generalizable framework based on LLM-SS that is modular by design, model-agnostic, domain-agnostic, and requires little to no human input.
Authors: Pingchuan Ma, Xiaopei Yang, Yusong Li, Ming Gui, Felix Krause, Johannes Schusterbauer, Bj\"orn Ommer
Abstract: Explicitly disentangling style and content in vision models remains challenging due to their semantic overlap and the subjectivity of human perception. Existing methods propose separation through generative or discriminative objectives, but they still face the inherent ambiguity of disentangling intertwined concepts. Instead, we ask: Can we bypass explicit disentanglement by learning to merge style and content invertibly, allowing separation to emerge naturally? We propose SCFlow, a flow-matching framework that learns bidirectional mappings between entangled and disentangled representations. Our approach is built upon three key insights: 1) Training solely to merge style and content, a well-defined task, enables invertible disentanglement without explicit supervision; 2) flow matching bridges on arbitrary distributions, avoiding the restrictive Gaussian priors of diffusion models and normalizing flows; and 3) a synthetic dataset of 510,000 samples (51 styles $\times$ 10,000 content samples) was curated to simulate disentanglement through systematic style-content pairing. Beyond controllable generation tasks, we demonstrate that SCFlow generalizes to ImageNet-1k and WikiArt in zero-shot settings and achieves competitive performance, highlighting that disentanglement naturally emerges from the invertible merging process.
Authors: Gang Yang
Abstract: Hyperspectral unmixing aims at estimating material signatures (known as endmembers) and the corresponding proportions (referred to abundances), which is a critical preprocessing step in various hyperspectral imagery applications. This study develops a novel approach called sparsity and total variation (TV) constrained multilayer linear unmixing (STVMLU) for hyperspectral imagery. Specifically, based on a multilayer matrix factorization model, to improve the accuracy of unmixing, a TV constraint is incorporated to consider adjacent spatial similarity. Additionally, a L1/2-norm sparse constraint is adopted to effectively characterize the sparsity of the abundance matrix. For optimizing the STVMLU model, the method of alternating direction method of multipliers (ADMM) is employed, which allows for the simultaneous extraction of endmembers and their corresponding abundance matrix. Experimental results illustrate the enhanced performance of the proposed STVMLU when compared to other algorithms.
Authors: Fei Shuang, Zixiong Wei, Kai Liu, Wei Gao, Poulumi Dey
Abstract: Machine learning interatomic potentials (MLIPs) enable accurate atomistic modelling, but reliable uncertainty quantification (UQ) remains elusive. In this study, we investigate two UQ strategies, ensemble learning and D-optimality, within the atomic cluster expansion framework. It is revealed that higher model accuracy strengthens the correlation between predicted uncertainties and actual errors and improves novelty detection, with D-optimality yielding more conservative estimates. Both methods deliver well calibrated uncertainties on homogeneous training sets, yet they underpredict errors and exhibit reduced novelty sensitivity on heterogeneous datasets. To address this limitation, we introduce clustering-enhanced local D-optimality, which partitions configuration space into clusters during training and applies D-optimality within each cluster. This approach substantially improves the detection of novel atomic environments in heterogeneous datasets. Our findings clarify the roles of model fidelity and data heterogeneity in UQ performance and provide a practical route to robust active learning and adaptive sampling strategies for MLIP development.
Authors: Futian Wang, Yuhan Qiao, Xiao Wang, Fuling Wang, Yuxiang Zhang, Dengdi Sun
Abstract: X-ray medical report generation is one of the important applications of artificial intelligence in healthcare. With the support of large foundation models, the quality of medical report generation has significantly improved. However, challenges such as hallucination and weak disease diagnostic capability still persist. In this paper, we first construct a large-scale multi-modal medical knowledge graph (termed M3KG) based on the ground truth medical report using the GPT-4o. It contains 2477 entities, 3 kinds of relations, 37424 triples, and 6943 disease-aware vision tokens for the CheXpert Plus dataset. Then, we sample it to obtain multi-granularity semantic graphs and use an R-GCN encoder for feature extraction. For the input X-ray image, we adopt the Swin-Transformer to extract the vision features and interact with the knowledge using cross-attention. The vision tokens are fed into a Q-former and retrieved the disease-aware vision tokens using another cross-attention. Finally, we adopt the large language model to map the semantic knowledge graph, input X-ray image, and disease-aware vision tokens into language descriptions. Extensive experiments on multiple datasets fully validated the effectiveness of our proposed knowledge graph and X-ray report generation framework. The source code of this paper will be released on https://github.com/Event-AHU/Medical_Image_Analysis.
Authors: Bojan Deraji\'c, Mohamed-Khalil Bouzidi, Sebastian Bernhard, Wolfgang H\"onig
Abstract: In this paper, we propose a hybrid MPC local planner that uses a learning-based approximation of a time-varying safe set, derived from local observations and applied as the MPC terminal constraint. This set can be represented as a zero-superlevel set of the value function computed via Hamilton-Jacobi (HJ) reachability analysis, which is infeasible in real-time. We exploit the property that the HJ value function can be expressed as a difference of the corresponding signed distance function (SDF) and a non-negative residual function. The residual component is modeled as a neural network with non-negative output and subtracted from the computed SDF, resulting in a real-time value function estimate that is at least as safe as the SDF by design. Additionally, we parametrize the neural residual by a hypernetwork to improve real-time performance and generalization properties. The proposed method is compared with three state-of-the-art methods in simulations and hardware experiments, achieving up to 30\% higher success rates compared to the best baseline while requiring a similar computational effort and producing high-quality (low travel-time) solutions.
Authors: Erico Souza Teixeira, Lucas Barros Fernandes, Yara Rodrigues In\'acio
Abstract: Binding energy is a fundamental thermodynamic property that governs molecular interactions, playing a crucial role in fields such as healthcare and the natural sciences. It is particularly relevant in drug development, vaccine design, and other biomedical applications. Over the years, various methods have been developed to estimate protein binding energy, ranging from experimental techniques to computational approaches, with machine learning making significant contributions to this field. Although classical computing has demonstrated strong results in constructing predictive models, the variation of quantum computing for machine learning has emerged as a promising alternative. Quantum neural networks (QNNs) have gained traction as a research focus, raising the question of their potential advantages in predicting binding energies. To investigate this potential, this study explored the feasibility of QNNs for this task by proposing thirty variations of multilayer perceptron-based quantum neural networks. These variations span three distinct architectures, each incorporating ten different quantum circuits to configure their quantum layers. The performance of these quantum models was compared with that of a state-of-the-art classical multilayer perceptron-based artificial neural network, evaluating both accuracy and training time. A primary dataset was used for training, while two additional datasets containing entirely unseen samples were employed for testing. Results indicate that the quantum models achieved approximately 20% higher accuracy on one unseen dataset, although their accuracy was lower on the other datasets. Notably, quantum models exhibited training times several orders of magnitude shorter than their classical counterparts, highlighting their potential for efficient protein binding energy prediction.
Authors: Rita Gonz\'alez-M\'arquez, Philipp Berens, Dmitry Kobak
Abstract: Text embeddings, i.e. vector representations of entire texts, play an important role in many NLP applications, such as retrieval-augmented generation, sentiment analysis, clustering, or visualizing collections of texts for data exploration. Currently, top-performing embedding models are derived from pre-trained language models via extensive supervised fine-tuning using curated text pairs. This contrasts with computer vision, where self-supervised training based on data augmentations has demonstrated remarkable success. Here we systematically compare the two most well-known augmentation strategies for positive pair generation in contrastive learning of text embeddings. We assess embedding quality on MTEB and additional in-domain evaluations and show that cropping augmentation strongly outperforms the dropout-based approach. We find that on out-of-domain data, the quality of resulting embeddings is below the supervised SOTA models, but for in-domain data, self-supervised fine-tuning produces high-quality text embeddings after very short fine-tuning, sometimes only marginally below the supervised SOTA. Finally, we show that representation quality increases towards the last transformer layers, which undergo the largest change during fine-tuning; and that fine-tuning only those last layers is sufficient to reach similar embedding quality.
Authors: Yuanpeng Li, Qi Long, Zhiyuan Yao, Jian Xu, Lintao Xie, Xu He, Lu Geng, Xin Han, Yueyan Chen, Wenbo Duan
Abstract: As enterprise codebases continue to grow in scale and complexity, the volume of lint errors far exceeds engineers' manual remediation capacity, leading to continuous accumulation of technical debt and hindered development efficiency. This paper presents BitsAI-Fix, an automated lint error remediation workflow based on Large Language Models (LLMs), designed to address this critical challenge in industrial-scale environments. BitsAI-Fix employs tree-sitter for context expansion and generates search-and-replace format patches through specially trained LLMs, followed by lint scan re-verification to output final remediation results. Additionally, our approach introduces an innovative progressive reinforcement learning (RL) training strategy that can automatically acquire verifiable training data during the project cold-start phase and continuously iterate the model by collecting online samples through feedback after system deployment. Furthermore, we designed a targeted rule-based reward mechanism that combines format rewards and correctness rewards while penalizing redundant modifications. We also propose a "code diff matching" methodology to continuously track online effectiveness. In production deployment at ByteDance, our solution has supported over 5,000 engineers, resolved more than 12,000 static analysis issues, achieved approximately 85% remediation accuracy, with around 1,000 weekly active adopters. This work demonstrates the practical feasibility of LLM-based code remediation solutions in enterprise environments and serves as a reference for automated code fix in large-scale industrial scenarios.
Authors: Marta Moscati, Shah Nawaz, Markus Schedl
Abstract: Recommender Systems (RS) often rely on representations of users and items in a joint embedding space and on a similarity metric to compute relevance scores. In modern RS, the modules to obtain user and item representations consist of two distinct and separate neural networks (NN). In multimodal representation learning, weight sharing has been proven effective in reducing the distance between multiple modalities of a same item. Inspired by these approaches, we propose a novel RS that leverages weight sharing between the user and item NN modules used to obtain the latent representations in the shared embedding space. The proposed framework consists of a single Collaborative Branch for Recommendation (CoBraR). We evaluate CoBraR by means of quantitative experiments on e-commerce and movie recommendation. Our experiments show that by reducing the number of parameters and improving beyond-accuracy aspects without compromising accuracy, CoBraR has the potential to be applied and extended for real-world scenarios.
Authors: Md Rakibul Hasan, Md Zakir Hossain, Aneesh Krishna, Shafin Rahman, Tom Gedeon
Abstract: Supervised learning for empathy regression is challenged by noisy self-reported empathy scores. While many algorithms have been proposed for learning with noisy labels in textual classification problems, the regression counterpart is relatively under-explored. We propose UPLME, an uncertainty-aware probabilistic language modelling framework to capture label noise in the regression setting of empathy detection. UPLME includes a probabilistic language model that predicts both empathy score and heteroscedastic uncertainty and is trained using Bayesian concepts with variational model ensembling. We further introduce two novel loss components: one penalises degenerate Uncertainty Quantification (UQ), and another enforces the similarity between the input pairs on which we predict empathy. UPLME provides state-of-the-art performance (Pearson Correlation Coefficient: $0.558\rightarrow0.580$ and $0.629\rightarrow0.634$) in terms of the performance reported in the literature in two public benchmarks, having label noise. Through synthetic label noise injection, we show that UPLME is effective in separating noisy and clean samples based on the predicted uncertainty. UPLME further outperform (Calibration error: $0.571\rightarrow0.376$) a recent variational model ensembling-based UQ method designed for regression problems.
Authors: Mohsen Romeshkani, J\"urgen M\"uller, Sahar Ebadi, Alexey Kupriyanov, Annike Knabe, Nina Fletling, Manuel Schilling
Abstract: Accurate and continuous monitoring of Earth's gravity field is essential for tracking mass redistribution processes linked to climate variability, hydrological cycles, and geodynamic phenomena. While the GRACE and GRACE Follow-On (GRACE-FO) missions have set the benchmark for satellite gravimetry using low-low satellite to satellite tracking (LL-SST), the precision of gravity field recovery still strongly depends on the quality of accelerometer (ACC) performance and the continuity of ACC data. Traditional electrostatic accelerometers (EA) face limitations that can hinder mission outcomes, prompting exploration of advanced sensor technologies and data recovery techniques. This study presents a systematic evaluation of accelerometer data transplantation using novel accelerometer configurations, including Cold Atom Interferometry (CAI) accelerometers and hybrid EA-CAI setups, and applying both analytical and machine learning-based methods. Using comprehensive closed-loop LL-SST simulations, we compare four scenarios ranging from the conventional EA-only setup to ideal dual hybrid configurations, with a particular focus on the performance of transplant-based approaches using different neural network approaches. Our results show that the dual hybrid configuration provides the most accurate gravity field retrieval. However, the transplant-based hybrid setup, especially when supported by machine learning, emerges as a robust and cost-effective alternative, achieving comparable performance with minimal extra hardware. These findings highlight the promise of combining quantum sensor technology and data-driven transplantation for future satellite gravimetry missions, paving the way for improved global monitoring of Earth's dynamic gravity field.
Authors: Stefan Brandst\"atter, Maximilian K\"oller, Philipp Seeb\"ock, Alissa Blessing, Felicitas Oberndorfer, Svitlana Pochepnia, Helmut Prosch, Georg Langs
Abstract: In histopathology, tissue samples are often larger than a standard microscope slide, making stitching of multiple fragments necessary to process entire structures such as tumors. Automated stitching is a prerequisite for scaling analysis, but is challenging due to possible tissue loss during preparation, inhomogeneous morphological distortion, staining inconsistencies, missing regions due to misalignment on the slide, or frayed tissue edges. This limits state-of-the-art stitching methods using boundary shape matching algorithms to reconstruct artificial whole mount slides (WMS). Here, we introduce SemanticStitcher using latent feature representations derived from a visual histopathology foundation model to identify neighboring areas in different fragments. Robust pose estimation based on a large number of semantic matching candidates derives a mosaic of multiple fragments to form the WMS. Experiments on three different histopathology datasets demonstrate that SemanticStitcher yields robust WMS mosaicing and consistently outperforms the state of the art in correct boundary matches.
Authors: Ergi Tushe, Bilal Farooq
Abstract: The integration of Automated Delivery Robots (ADRs) into pedestrian-heavy urban spaces introduces unique challenges in terms of safe, efficient, and socially acceptable navigation. We develop the complete pipeline for a single vision sensor based multi-pedestrian detection and tracking, pose estimation, and monocular depth perception. Leveraging the real-world MOT17 dataset sequences, this study demonstrates how integrating human-pose estimation and depth cues enhances pedestrian trajectory prediction and identity maintenance, even under occlusions and dense crowds. Results show measurable improvements, including up to a 10% increase in identity preservation (IDF1), a 7% improvement in multiobject tracking accuracy (MOTA), and consistently high detection precision exceeding 85%, even in challenging scenarios. Notably, the system identifies vulnerable pedestrian groups supporting more socially aware and inclusive robot behaviour.
Authors: Zhanye Luo, Yuefeng Han, Xiufan Yu
Abstract: This paper studies the problem of dimension reduction, tailored to improving time series forecasting with high-dimensional predictors. We propose a novel Supervised Deep Dynamic Principal component analysis (SDDP) framework that incorporates the target variable and lagged observations into the factor extraction process. Assisted by a temporal neural network, we construct target-aware predictors by scaling the original predictors in a supervised manner, with larger weights assigned to predictors with stronger forecasting power. A principal component analysis is then performed on the target-aware predictors to extract the estimated SDDP factors. This supervised factor extraction not only improves predictive accuracy in the downstream forecasting task but also yields more interpretable and target-specific latent factors. Building upon SDDP, we propose a factor-augmented nonlinear dynamic forecasting model that unifies a broad family of factor-model-based forecasting approaches. To further demonstrate the broader applicability of SDDP, we extend our studies to a more challenging scenario when the predictors are only partially observable. We validate the empirical performance of the proposed method on several real-world public datasets. The results show that our algorithm achieves notable improvements in forecasting accuracy compared to state-of-the-art methods.
Authors: Iing Muttakhiroh, Thomas Fevens
Abstract: Large Language Models (LLMs) often suffer from performance degradation when faced with domain shifts, primarily due to catastrophic forgetting. In this work, we propose KILO (Knowledge-Instructed Learning for Continual Adaptation), a novel continual learning framework that integrates dynamic knowledge graphs with instruction tuning. By leveraging retrieved domain-specific knowledge as guidance during training, KILO enhances both adaptability to new domains and retention of previously acquired knowledge. We pretrain our model on WikiText-103 and evaluate sequential adaptation across four diverse target domains: BioASQ, SciQ, TweetEval, and MIND. Our experiments demonstrate that KILO consistently outperforms strong baselines, including continual fine-tuning, ERNIE 2.0, and CPT, in terms of backward transfer, forward transfer, F1 score, retention rate, and training efficiency. These results highlight the effectiveness of combining structured knowledge retrieval and instruction prompting to overcome domain shift challenges in continual learning scenarios.
Authors: Md Abrar Jahin, Shahriar Soudeep, M. F. Mridha, Nafiz Fahad, Md. Jakir Hossen
Abstract: Recent advancements in object detection rely on modular architectures with multi-scale fusion and attention mechanisms. However, static fusion heuristics and class-agnostic attention limit performance in dynamic scenes with occlusions, clutter, and class imbalance. We introduce Dynamic Class-Aware Fusion Network (DyCAF-Net) that addresses these challenges through three innovations: (1) an input-conditioned equilibrium-based neck that iteratively refines multi-scale features via implicit fixed-point modeling, (2) a dual dynamic attention mechanism that adaptively recalibrates channel and spatial responses using input- and class-dependent cues, and (3) class-aware feature adaptation that modulates features to prioritize discriminative regions for rare classes. Through comprehensive ablation studies with YOLOv8 and related architectures, alongside benchmarking against nine state-of-the-art baselines, DyCAF-Net achieves significant improvements in precision, mAP@50, and mAP@50-95 across 13 diverse benchmarks, including occlusion-heavy and long-tailed datasets. The framework maintains computational efficiency ($\sim$11.1M parameters) and competitive inference speeds, while its adaptability to scale variance, semantic overlaps, and class imbalance positions it as a robust solution for real-world detection tasks in medical imaging, surveillance, and autonomous systems.
Authors: Soumik Dey, Benjamin Braun, Naveen Ravipati, Hansi Wu, Binbin Li
Abstract: Sellers at eBay are recommended keyphrases to bid on to enhance the performance of their advertising campaigns. The relevance of these keyphrases is crucial in avoiding the overcrowding of search systems with irrelevant items and maintaining a positive seller perception. It is essential that keyphrase recommendations align with both seller and Search judgments regarding auctions. Due to the difficulty in procuring negative human judgment at scale, employing LLM-as-a-judge to mimic seller judgment has been established as the norm in several studies. This study introduces a novel two-step LLM distillation process from a LLM-judge used to debias our Embedding Based Retrieval (EBR) model from the various biases that exist in click-data. We distill from an LLM teacher via a cross-encoder assistant into a bi-encoder student using a multi-task training approach, ultimately employing the student bi-encoder to retrieve relevant advertiser keyphrases. We show that integrating a knowledge distillation process from LLMs in a multi-task training setup enhances bi-encoder performance in retrieving relevant advertiser keyphrases at eBay.
Authors: Lei Qian, Wu Su, Yanqi Huang, Song Xi Chen
Abstract: We propose a Likelihood Matching approach for training diffusion models by first establishing an equivalence between the likelihood of the target data distribution and a likelihood along the sample path of the reverse diffusion. To efficiently compute the reverse sample likelihood, a quasi-likelihood is considered to approximate each reverse transition density by a Gaussian distribution with matched conditional mean and covariance, respectively. The score and Hessian functions for the diffusion generation are estimated by maximizing the quasi-likelihood, ensuring a consistent matching of both the first two transitional moments between every two time points. A stochastic sampler is introduced to facilitate computation that leverages on both the estimated score and Hessian information. We establish consistency of the quasi-maximum likelihood estimation, and provide non-asymptotic convergence guarantees for the proposed sampler, quantifying the rates of the approximation errors due to the score and Hessian estimation, dimensionality, and the number of diffusion steps. Empirical and simulation evaluations demonstrate the effectiveness of the proposed Likelihood Matching and validate the theoretical results.
Authors: Akshay L Chandra, Iman Nematollahi, Chenguang Huang, Tim Welschehold, Wolfram Burgard, Abhinav Valada
Abstract: Fine-tuning diffusion policies with reinforcement learning (RL) presents significant challenges. The long denoising sequence for each action prediction impedes effective reward propagation. Moreover, standard RL methods require millions of real-world interactions, posing a major bottleneck for practical fine-tuning. Although prior work frames the denoising process in diffusion policies as a Markov Decision Process to enable RL-based updates, its strong dependence on environment interaction remains highly inefficient. To bridge this gap, we introduce DiWA, a novel framework that leverages a world model for fine-tuning diffusion-based robotic skills entirely offline with reinforcement learning. Unlike model-free approaches that require millions of environment interactions to fine-tune a repertoire of robot skills, DiWA achieves effective adaptation using a world model trained once on a few hundred thousand offline play interactions. This results in dramatically improved sample efficiency, making the approach significantly more practical and safer for real-world robot learning. On the challenging CALVIN benchmark, DiWA improves performance across eight tasks using only offline adaptation, while requiring orders of magnitude fewer physical interactions than model-free baselines. To our knowledge, this is the first demonstration of fine-tuning diffusion policies for real-world robotic skills using an offline world model. We make the code publicly available at https://diwa.cs.uni-freiburg.de.
Authors: Fernando F. Granado, Davi A. Bezerra, Iuri Queiroz, Nathan Oliveira, Pedro Fernandes, Bruno Schock
Abstract: Food delivery platforms face the challenge of helping users navigate vast catalogs of restaurants and dishes to find meals they truly enjoy. This paper presents RED, an automated recommendation system designed for iFood, Latin America's largest on-demand food delivery platform, to personalize the selection of curated food collections displayed to millions of users. Our approach employs a LightGBM classifier that scores collections based on three feature groups: collection characteristics, user-collection similarity, and contextual information. To address the cold-start problem of recommending newly created collections, we develop content-based representations using item embeddings and implement monotonicity constraints to improve generalization. We tackle data scarcity by bootstrapping from category carousel interactions and address visibility bias through unbiased sampling of impressions and purchases in production. The system demonstrates significant real-world impact through extensive A/B testing with 5-10% of iFood's user base. Online results of our A/B tests add up to 97% improvement in Card Conversion Rate and 1.4% increase in overall App Conversion Rate compared to popularity-based baselines. Notably, our offline accuracy metrics strongly correlate with online performance, enabling reliable impact prediction before deployment. To our knowledge, this is the first work to detail large-scale recommendation of curated food collections in a dynamic commercial environment.
Authors: Abhishek Vijaya Kumar, Eric Ding, Arjun Devraj, Rachee Singh
Abstract: We optically interconnect accelerator chips (e.g., GPUs, TPUs) within compute servers using newly viable programmable chip-to-chip photonic fabrics. In contrast, today, commercial multi-accelerator compute servers that are workhorses of ML, use electrical interconnects to network accelerator chips in the server. However, recent trends have shown an interconnect bandwidth wall caused by accelerator FLOPS scaling at a faster rate than the bandwidth of the interconnect between accelerators in the same server. This has led to under-utilization and idling of GPU resources in cloud datacenters. We develop Morphlux, a server-scale programmable photonic fabric, to interconnect accelerators within servers. We show that augmenting state-of-the-art photonic ML-centric datacenters with Morphlux can improve the bandwidth of tenant compute allocations by up to 66% and reduce compute fragmentation by up to 70%. We develop a novel end-to-end hardware prototype of Morphlux to demonstrate these performance benefits, which translate to 1.72x improvement in training throughput of ML models. By rapidly programming the server-scale fabric in our hardware testbed, Morphlux can logically replace a failed accelerator chip in 1.2 seconds.
Authors: Stella Bitchebe, Oana Balmau
Abstract: A large body of research has employed Machine Learning (ML) models to develop learned operating systems (OSes) and kernels. The latter dynamically adapts to the job load and dynamically adjusts resources (CPU, IO, memory, network bandwidth) allocation to respond to the actual user demand. What this work has in common is that it utilizes ML to improve kernel decisions. To this day, and to the best of our knowledge, no work has taken the opposite direction, i.e., using OS to improve ML. While some work proposes applying system-level optimizations to ML algorithms, they do not tailor the OS to adapt to the ML context. To address this limitation, we take an orthogonal approach in this paper by leveraging the OS to enhance the performance of ML models and algorithms. We explore the path towards an ML-specialized OS, MaLV-OS. MaLV-OS rethinks the OS architecture to make it specifically tailored to ML workloads, especially in virtualized clouds, which are now widely used to run ML applications. MaLV-OS envisioned architecture includes (1) a micro-kernel, Micro-LAKE, which allows kernel space applications to use the GPU, and (2) an MLaaS (ML as a Service) subsystem that gathers ML models to help Micro-LAKE with memory management and CPU scheduling. MaLV-OS architecture also offloads system-sensitive parts of the models to the OS, to lighten the model complexity and programming, and speed up its execution. Finally, MaLV-OS integrates an open-source GPU virtualization software, merged directly into the hypervisor. For more flexibility, MaLV-OS vision is to enable the virtual machine to dynamically select MLaaS policies that can improve the performance of the model the user is running. Because MLaaS is designed as loadable kernel modules, the MaLV-OS architecture enables the dynamic addition of new capabilities to the MLaaS subsystem.
Authors: Yangtian Zi, Harshitha Menon, Arjun Guha
Abstract: State-of-the-art Large Language Models (LLMs) achieve high pass@1 on general benchmarks like HumanEval but underperform on specialized suites such as ParEval. Is this due to LLMs missing domain knowledge or insufficient prompt detail is given? To answer this, we introduce PartialOrderEval, which augments any code generation benchmark with a partial order of prompts from minimal to maximally detailed. Applying it to HumanEval and both serial and OpenMP subsets of ParEval, we measure how pass@1 scales with prompt specificity. Our experiments with Llama-3.x and Qwen2.5-Coder demonstrate varying degrees of prompt sensitivity across different tasks, and a qualitative analysis highlights explicit I/O specifications, edge-case handling, and stepwise breakdowns as the key drivers of prompt detail improvement.
Authors: Xufang Luo, Yuge Zhang, Zhiyuan He, Zilong Wang, Siyun Zhao, Dongsheng Li, Luna K. Qiu, Yuqing Yang
Abstract: We present Agent Lightning, a flexible and extensible framework that enables Reinforcement Learning (RL)-based training of Large Language Models (LLMs) for any AI agent. Unlike existing methods that tightly couple RL training with agent or rely on sequence concatenation with masking, Agent Lightning achieves complete decoupling between agent execution and training, allowing seamless integration with existing agents developed via diverse ways (e.g., using frameworks like LangChain, OpenAI Agents SDK, AutoGen, and building from scratch) with almost ZERO code modifications. By formulating agent execution as Markov decision process, we define an unified data interface and propose a hierarchical RL algorithm, LightningRL, which contains a credit assignment module, allowing us to decompose trajectories generated by ANY agents into training transition. This enables RL to handle complex interaction logic, such as multi-agent scenarios and dynamic workflows. For the system design, we introduce a Training-Agent Disaggregation architecture, and brings agent observability frameworks into agent runtime, providing a standardized agent finetuning interface. Experiments across text-to-SQL, retrieval-augmented generation, and math tool-use tasks demonstrate stable, continuous improvements, showcasing the framework's potential for real-world agent training and deployment.
Authors: Shreya Meel, Mohamed Nomeir, Pasan Dissanayake, Sanghamitra Dutta, Sennur Ulukus
Abstract: Transparency and explainability are two important aspects to be considered when employing black-box machine learning models in high-stake applications. Providing counterfactual explanations is one way of catering this requirement. However, this also poses a threat to the privacy of the institution that is providing the explanation, as well as the user who is requesting it. In this work, we are primarily concerned with the user's privacy who wants to retrieve a counterfactual instance, without revealing their feature vector to the institution. Our framework retrieves the exact nearest neighbor counterfactual explanation from a database of accepted points while achieving perfect, information-theoretic, privacy for the user. First, we introduce the problem of private counterfactual retrieval (PCR) and propose a baseline PCR scheme that keeps the user's feature vector information-theoretically private from the institution. Building on this, we propose two other schemes that reduce the amount of information leaked about the institution database to the user, compared to the baseline scheme. Second, we relax the assumption of mutability of all features, and consider the setting of immutable PCR (I-PCR). Here, the user retrieves the nearest counterfactual without altering a private subset of their features, which constitutes the immutable set, while keeping their feature vector and immutable set private from the institution. For this, we propose two schemes that preserve the user's privacy information-theoretically, but ensure varying degrees of database privacy. Third, we extend our PCR and I-PCR schemes to incorporate user's preference on transforming their attributes, so that a more actionable explanation can be received. Finally, we present numerical results to support our theoretical findings, and compare the database leakage of the proposed schemes.
Authors: G\'erard Ben Arous, Murat A. Erdogdu, N. Mert Vural, Denny Wu
Abstract: We study the optimization and sample complexity of gradient-based training of a two-layer neural network with quadratic activation function in the high-dimensional regime, where the data is generated as $y \propto \sum_{j=1}^{r}\lambda_j \sigma\left(\langle \boldsymbol{\theta_j}, \boldsymbol{x}\rangle\right), \boldsymbol{x} \sim N(0,\boldsymbol{I}_d)$, $\sigma$ is the 2nd Hermite polynomial, and $\lbrace\boldsymbol{\theta}_j \rbrace_{j=1}^{r} \subset \mathbb{R}^d$ are orthonormal signal directions. We consider the extensive-width regime $r \asymp d^\beta$ for $\beta \in [0, 1)$, and assume a power-law decay on the (non-negative) second-layer coefficients $\lambda_j\asymp j^{-\alpha}$ for $\alpha \geq 0$. We present a sharp analysis of the SGD dynamics in the feature learning regime, for both the population limit and the finite-sample (online) discretization, and derive scaling laws for the prediction risk that highlight the power-law dependencies on the optimization time, sample size, and model width. Our analysis combines a precise characterization of the associated matrix Riccati differential equation with novel matrix monotonicity arguments to establish convergence guarantees for the infinite-dimensional effective dynamics.
Authors: Paz Fink Shustin, Shashanka Ubaru, Ma{\l}gorzata J. Zimo\'n, Songtao Lu, Vasileios Kalantzis, Lior Horesh, Haim Avron
Abstract: Learning data representations under uncertainty is an important task that emerges in numerous scientific computing and data analysis applications. However, uncertainty quantification techniques are computationally intensive and become prohibitively expensive for high-dimensional data. In this study, we introduce a dimensionality reduction surrogate modeling (DRSM) approach for representation learning and uncertainty quantification that aims to deal with data of moderate to high dimensions. The approach involves a two-stage learning process: 1) employing a variational autoencoder to learn a low-dimensional representation of the input data distribution; and 2) harnessing polynomial chaos expansion (PCE) formulation to map the low dimensional distribution to the output target. The model enables us to (a) capture the system dynamics efficiently in the low-dimensional latent space, (b) learn under uncertainty, a representation of the data and a mapping between input and output distributions, (c) estimate this uncertainty in the high-dimensional data system, and (d) match high-order moments of the output distribution; without any prior statistical assumptions on the data. Numerical results are presented to illustrate the performance of the proposed method.
Authors: Ziji Shi, Le Jiang, Ang Wang, Jie Zhang, Chencan Wu, Yong Li, Xiaokui Xiao, Wei Lin, Jialin Li
Abstract: Tensor parallelism is an essential technique for distributed training of large neural networks. However, automatically determining an optimal tensor parallel strategy is challenging due to the gigantic search space, which grows exponentially with model size and tensor dimension. This prohibits the adoption of auto-parallel systems on larger models. We observe that neural networks usually contain repeated substructures, and build an automatic parallelism framework named TAPAS that eliminates redundant search efforts. TAPAS employs a divide-and-conquer approach that efficiently folds the search space by identifying those unique substructures. As a result, it runs at sub-linear complexity concerning the model size, making it a scalable solution for training large-scale networks. Our evaluations demonstrate that TAPAS outperforms the state-of-the-art automatic parallelism frameworks by up to $160\times$ in search speed on a wide range of models, and the performance of derived strategies is competitive or even better compared with the expert-engineered Megatron-LM library.
Authors: Hao Di, Yi Yang, Haishan Ye, Xiangyu Chang
Abstract: Personalization aims to characterize individual preferences and is widely applied across many fields. However, conventional personalized methods operate in a centralized manner, potentially exposing raw data when pooling individual information. In this paper, with privacy considerations, we develop a flexible and interpretable personalized framework within the paradigm of federated learning, called \texttt{PPFL} (Population Personalized Federated Learning). By leveraging ``canonical models" to capture fundamental characteristics of a heterogeneous population and employing ``membership vectors" to reveal clients' preferences, \texttt{PPFL} models heterogeneity as clients' varying preferences for these characteristics. This approach provides substantial insights into client characteristics, which are lacking in existing Personalized Federated Learning (PFL) methods. Furthermore, we explore the relationship between \texttt{PPFL} and three main branches of PFL methods: clustered FL, multi-task PFL, and decoupling PFL, and demonstrate the advantages of \texttt{PPFL}. To solve \texttt{PPFL} (a non-convex optimization problem with linear constraints), we propose a novel random block coordinate descent algorithm and establish its convergence properties. We conduct experiments on both pathological and practical data sets, and the results validate the effectiveness of \texttt{PPFL}.
Authors: Lukas Koller, Tobias Ladner, Matthias Althoff
Abstract: Neural networks are vulnerable to adversarial attacks, i.e., small input perturbations can significantly affect the outputs of a neural network. Therefore, to ensure safety of neural networks in safety-critical environments, the robustness of a neural network must be formally verified against input perturbations, e.g., from noisy sensors. To improve the robustness of neural networks and thus simplify the formal verification, we present a novel set-based training procedure in which we compute the set of possible outputs given the set of possible inputs and compute for the first time a gradient set, i.e., each possible output has a different gradient. Therefore, we can directly reduce the size of the output enclosure by choosing gradients toward its center. Small output enclosures increase the robustness of a neural network and, at the same time, simplify its formal verification. The latter benefit is due to the fact that a larger size of propagated sets increases the conservatism of most verification methods. Our extensive evaluation demonstrates that set-based training produces robust neural networks with competitive performance, which can be verified using fast (polynomial-time) verification algorithms due to the reduced output set.
Authors: Leon G\"otz, Marcel Kollovieh, Stephan G\"unnemann, Leo Schwinn
Abstract: Despite recent advances in subquadratic attention mechanisms or state-space models, processing long token sequences still imposes significant computational requirements. Token merging has emerged as a solution to increase computational efficiency in computer vision architectures. In this work, we perform the first investigations of token merging in time series analysis on both transformers and state-space models. We further introduce local merging, a domain-specific token merging algorithm that selectively combines tokens within a local neighborhood, achieving two major benefits: a) Local merging can adjust its computational complexity from quadratic to linear based on the neighborhood size to effectively scale to long sequences; b) Local merging is the first causal merging scheme enabling token merging in transformer decoders. Further, we identify spectral properties of the input data that reliably predict the potential benefits of local merging without requiring evaluation on downstream tasks. Our comprehensive empirical evaluation demonstrates that local merging offers substantial efficiency gains with minimal impact on accuracy, achieving up to 5400% acceleration on the recently proposed Chronos foundation model.
Authors: Michael Lingzhi Li, Shixiang Zhu
Abstract: Operational decisions in healthcare, logistics, and public policy increasingly involve algorithms that recommend candidate solutions, such as treatment plans, delivery routes, or policy options, while leaving the final choice to human decision-makers. For instance, school districts use algorithms to design bus routes, but administrators make the final call given community feedback. In these settings, decision quality depends not on a single algorithmic ``optimum'', but on whether the portfolio of recommendations contains at least one option the human ultimately deems desirable. We propose generative curation, a framework that optimally generates recommendation sets when desirability depends on both observable objectives and unobserved qualitative considerations. Instead of a fixed solution, generative curation learns a distribution over solutions designed to maximize the expected desirability of the best option within a manageable portfolio. Our analysis identifies a trade-off between quantitative quality and qualitative diversity, formalized through a novel diversity metric derived from the reformulated objective. We implement the framework using a generative neural network and a sequential optimization method, and show in synthetic and real-world studies that it consistently reduces expected regret compared to existing benchmarks. Our framework provides decision-makers with a principled way to design algorithms that complement, rather than replace, human judgment. By generating portfolios of diverse yet high-quality options, decision-support tools can better accommodate unmodeled factors such as stakeholder preferences, political feasibility, or community acceptance. More broadly, the framework enables organizations to operationalize human-centered decision-making at scale, ensuring that algorithmic recommendations remain useful even when objectives are incomplete or evolving.
Authors: Siyi Wang, Zifan Wang, Karl Henrik Johansson, Sandra Hirche
Abstract: In real-world scenarios, risk-averse learning is valuable for mitigating potential adverse outcomes. However, the delayed feedback makes it challenging to assess and manage risk effectively. In this paper, we investigate risk-averse learning using Conditional Value at Risk (CVaR) as risk measure, while incorporating feedback with random but bounded delays. We develop two risk-averse learning algorithms that rely on one-point and two-point zeroth-order optimization approaches, respectively. The dynamic regrets of the algorithms are analyzed in terms of the cumulative delay and the number of total samplings. In the absence of delay, the regret bounds match the established bounds of zeroth-order stochastic gradient methods for risk-averse learning. Furthermore, the two-point risk-averse learning outperforms the one-point algorithm by achieving a smaller regret bound. We provide numerical experiments on a dynamic pricing problem to demonstrate the performance of the algorithms.
Authors: T. E. R\"ober, R. Goedhart, S. \.I. Birbil
Abstract: Explainable AI (XAI) holds the promise of advancing the implementation and adoption of AI-based tools in practice, especially in high-stakes environments like healthcare. However, most of the current research lacks input from end users, and therefore their practical value is limited. To address this, we conducted semi-structured interviews with clinicians to discuss their thoughts, hopes, and concerns. Clinicians from our sample generally think positively about developing AI-based tools for clinical practice, but they have concerns about how these will fit into their workflow and how it will impact clinician-patient relations. We further identify training of clinicians on AI as a crucial factor for the success of AI in healthcare and highlight aspects clinicians are looking for in (X)AI-based tools. In contrast to other studies, we take on a holistic and exploratory perspective to identify general requirements for (X)AI products for healthcare before moving on to testing specific tools.
Authors: Jiaxi Li, Lu Yin, Xilu Wang
Abstract: The integration of Large Language Models (LLMs) into autonomous driving systems offers promising enhancements in environmental understanding and decision-making. However, the substantial computational demands of deploying LLMs locally on vehicles render this approach unfeasible for real-world automotive applications. To address this challenge, we introduce OWLed, the Outlier-Weighed Layerwise Pruning for Efficient Autonomous Driving Framework that leverages outlier-weighted layerwise sparsity for model compression. Our method assigns non-uniform sparsity ratios to different layers based on the distribution of outlier features, significantly reducing the model size without the need for fine-tuning. To ensure the compressed model adapts well to autonomous driving tasks, we incorporate driving environment data into both the calibration and pruning processes. Our empirical studies reveal that the encoder component is more sensitive to pruning than the LLM, highlighting its critical role in the system. Experimental results demonstrate that OWLed outperforms existing methods in perception, action prediction, and language understanding while substantially lowering computational requirements. These findings underscore the potential of combining advanced pruning techniques with LLMs to develop efficient and robust autonomous driving systems capable of handling complex scenarios. Code will be made publicly available.
Authors: Songning Lai, Mingqian Liao, Zhangyi Hu, Jiayu Yang, Wenshuo Chen, Hongru Xiao, Jianheng Tang, Haicheng Liao, Yutao Yue
Abstract: Concept Bottleneck Models (CBMs) enhance the interpretability of AI systems, particularly by bridging visual input with human-understandable concepts, effectively acting as a form of multimodal interpretability model. However, existing CBMs typically assume static datasets, which fundamentally limits their adaptability to real-world, continuously evolving multimodal data streams. To address this, we define a novel continual learning task for CBMs: simultaneously handling concept-incremental and class-incremental learning. This task requires models to continuously acquire new concepts (often representing cross-modal attributes) and classes while robustly preserving previously learned knowledge. To tackle this challenging problem, we propose CONceptual Continual Incremental Learning (CONCIL), a novel framework that fundamentally re-imagines concept and decision layer updates as linear regression problems. This reformulation eliminates the need for gradient-based optimization, thereby effectively preventing catastrophic forgetting. Crucially, CONCIL relies solely on recursive matrix operations, rendering it highly computationally efficient and well-suited for real-time and large-scale multimodal data applications. Experimental results compellingly demonstrate that CONCIL achieves "absolute knowledge memory" and significantly surpasses the performance of traditional CBM methods in both concept- and class-incremental settings, thus establishing a new paradigm for continual learning in CBMs, particularly valuable for dynamic multimodal understanding.
Authors: Luca Colombo, Fabrizio Pittorino, Manuel Roveri
Abstract: Binary Neural Networks (BNNs) significantly reduce computational complexity and memory usage in machine and deep learning by representing weights and activations with just one bit. However, most existing training algorithms for BNNs rely on quantization-aware floating-point Stochastic Gradient Descent (SGD), limiting the full exploitation of binary operations to the inference phase only. In this work, we propose, for the first time, a fully binary and gradient-free training algorithm for multi-layer BNNs, eliminating the need for back-propagated floating-point gradients. Specifically, the proposed algorithm relies on local binary error signals and binary weight updates, employing integer-valued hidden weights that serve as a synaptic metaplasticity mechanism, thereby enhancing its neurobiological plausibility. Our proposed solution enables the training of binary multi-layer perceptrons by using exclusively XNOR, Popcount, and increment/decrement operations. Experimental results on multi-class classification benchmarks show test accuracy improvements of up to +35.47% over the only existing fully binary single-layer state-of-the-art solution. Compared to full-precision SGD, our solution improves test accuracy by up to +35.30% under the same total memory demand, while also reducing computational cost by two to three orders of magnitude in terms of the total number of Boolean gates. The proposed algorithm is made available to the scientific community as a public repository.
Authors: Jacob Adamczyk, Volodymyr Makarenko, Stas Tiomkin, Rahul V. Kulkarni
Abstract: The average-reward formulation of reinforcement learning (RL) has drawn increased interest in recent years for its ability to solve temporally-extended problems without relying on discounting. Meanwhile, in the discounted setting, algorithms with entropy regularization have been developed, leading to improvements over deterministic methods. Despite the distinct benefits of these approaches, deep RL algorithms for the entropy-regularized average-reward objective have not been developed. While policy-gradient based approaches have recently been presented for the average-reward literature, the corresponding actor-critic framework remains less explored. In this paper, we introduce an average-reward soft actor-critic algorithm to address these gaps in the field. We validate our method by comparing with existing average-reward algorithms on standard RL benchmarks, achieving superior performance for the average-reward criterion.
Authors: F. S. Pezzicoli, V. Ros, F. P. Landes, M. Baity-Jesi
Abstract: Class imbalance (CI) is a longstanding problem in machine learning, slowing down training and reducing performances. Although empirical remedies exist, it is often unclear which ones work best and when, due to the lack of an overarching theory. We address a common case of imbalance, that of anomaly (or outlier) detection. We provide a theoretical framework to analyze, interpret and address CI. It is based on an exact solution of the teacher-student perceptron model, through replica theory. Within this framework, one can distinguish several sources of CI: either intrinsic, train or test imbalance. Our analysis reveals that the optimal train imbalance is generally different from 50%, with a non trivial dependence on the intrinsic imbalance, the abundance of data and on the noise in the learning. Moreover, there is a crossover between a small noise training regime where results are independent of the noise level to a high noise regime where performances quickly degrade with noise. Our results challenge some of the conventional wisdom on CI and offer practical guidelines to address it.
Authors: Wenyun Li, Wenjie Huang, Chen Sun
Abstract: In many real-world scenarios, reward signal for agents are exceedingly sparse, making it challenging to learn an effective reward function for reward shaping. To address this issue, the proposed approach in this paper performs reward shaping not only by utilizing non-zero-reward transitions but also by employing the \emph{Semi-Supervised Learning} (SSL) technique combined with a novel data augmentation to learn trajectory space representations from the majority of transitions, {i.e}., zero-reward transitions, thereby improving the efficacy of reward shaping. Experimental results in Atari and robotic manipulation demonstrate that our method outperforms supervised-based approaches in reward inference, leading to higher agent scores. Notably, in more sparse-reward environments, our method achieves up to twice the peak scores compared to supervised baselines. The proposed double entropy data augmentation enhances performance, showcasing a 15.8\% increase in best score over other augmentation methods
Authors: Yang Zhang, Wenbo Yang, Jun Wang, Qiang Ma, Jie Xiong
Abstract: Accurately forecasting the impact of macroeconomic events is critical for investors and policymakers. Salient events like monetary policy decisions and employment reports often trigger market movements by shaping expectations of economic growth and risk, thereby establishing causal relationships between events and market behavior. Existing forecasting methods typically focus either on textual analysis or time-series modeling, but fail to capture the multi-modal nature of financial markets and the causal relationship between events and price movements. To address these gaps, we propose CAMEF (Causal-Augmented Multi-Modality Event-Driven Financial Forecasting), a multi-modality framework that effectively integrates textual and time-series data with a causal learning mechanism and an LLM-based counterfactual event augmentation technique for causal-enhanced financial forecasting. Our contributions include: (1) a multi-modal framework that captures causal relationships between policy texts and historical price data; (2) a new financial dataset with six types of macroeconomic releases from 2008 to April 2024, and high-frequency real trading data for five key U.S. financial assets; and (3) an LLM-based counterfactual event augmentation strategy. We compare CAMEF to state-of-the-art transformer-based time-series and multi-modal baselines, and perform ablation studies to validate the effectiveness of the causal learning mechanism and event types.
Authors: Quan Xiao, Hui Yuan, A F M Saif, Gaowen Liu, Ramana Kompella, Mengdi Wang, Tianyi Chen
Abstract: Diffusion models, which iteratively denoise data samples to synthesize high-quality outputs, have achieved empirical success across domains. However, optimizing these models for downstream tasks often involves nested bilevel structures, such as tuning hyperparameters for fine-tuning tasks or noise schedules in training dynamics, where traditional bilevel methods fail due to the infinite-dimensional probability space and prohibitive sampling costs. We formalize this challenge as a generative bilevel optimization problem and address two key scenarios: (1) fine-tuning pre-trained models via an inference-only lower-level solver paired with a sample-efficient gradient estimator for the upper level, and (2) training diffusion model from scratch with noise schedule optimization by reparameterizing the lower-level problem and designing a computationally tractable gradient estimator. Our first-order bilevel framework overcomes the incompatibility of conventional bilevel methods with diffusion processes, offering theoretical grounding and computational practicality. Experiments demonstrate that our method outperforms existing fine-tuning and hyperparameter search baselines.
Authors: Shuo Wang, Keke Gai, Jing Yu, Liehuang Zhu, Qi Wu
Abstract: Vertical Federated Learning (VFL) has garnered significant attention as a privacy-preserving machine learning framework for sample-aligned feature federation. However, traditional VFL approaches do not address the challenges of class and feature continual learning, resulting in catastrophic forgetting of knowledge from previous tasks. To address the above challenge, we propose a novel vertical federated continual learning method, named Vertical Federated Continual Learning via Evolving Prototype Knowledge (V-LETO), which primarily facilitates the transfer of knowledge from previous tasks through the evolution of prototypes. Specifically, we propose an evolving prototype knowledge method, enabling the global model to retain both previous and current task knowledge. Furthermore, we introduce a model optimization technique that mitigates the forgetting of previous task knowledge by restricting updates to specific parameters of the local model, thereby enhancing overall performance. Extensive experiments conducted in both CIL and FIL settings demonstrate that our method, V-LETO, outperforms the other state-of-the-art methods. For example, our method outperforms the state-of-the-art method by 10.39% and 35.15% for CIL and FIL tasks, respectively. Our code is available at https://anonymous.4open.science/r/V-LETO-0108/README.md.
URLs: https://anonymous.4open.science/r/V-LETO-0108/README.md.
Authors: Riccardo Ali, Francesco Caso, Christopher Irwin, Pietro Li\`o
Abstract: Transformer models map input token sequences to output token distributions, layer by layer. While most interpretability work focuses on internal latent representations, we study the evolution of these token-level distributions directly in vocabulary space. However, such distributions are high-dimensional and defined on an unordered support, making common descriptors like moments or cumulants ill-suited. We address this by computing the Shannon entropy of each intermediate predicted distribution, yielding one interpretable scalar per layer. The resulting sequence, the entropy profile, serves as a compact, information-theoretic signature of the model's computation. We introduce Entropy-Lens, a model-agnostic framework that extracts entropy profiles from frozen, off-the-shelf transformers. We show that these profiles (i) reveal family-specific computation patterns invariant under depth rescaling, (ii) are predictive of prompt type and task format, and (iii) correlate with output correctness. We further show that R\'enyi entropies yield similar results within a broad range of $\alpha$ values, justifying the use of Shannon entropy as a stable and principled summary. Our results hold across different transformers, without requiring gradients, fine-tuning, or access to model internals.
Authors: Shengbo Gong, Mohammad Hashemi, Juntong Ni, Carl Yang, Wei Jin
Abstract: The rapid growth of graph data creates significant scalability challenges as most graph algorithms scale quadratically with size. To mitigate these issues, Graph Condensation (GC) methods have been proposed to learn a small graph from a larger one, accelerating downstream tasks. However, existing approaches critically assume a static training set, which conflicts with the inherently dynamic and evolving nature of real-world graph data. This work introduces a novel framework for continual graph condensation, enabling efficient updates to the distilled graph that handle data streams without requiring costly retraining. This limitation leads to inefficiencies when condensing growing training sets. In this paper, we introduce GECC (\underline{G}raph \underline{E}volving \underline{C}lustering \underline{C}ondensation), a scalable graph condensation method designed to handle large-scale and evolving graph data. GECC employs a traceable and efficient approach by performing class-wise clustering on aggregated features. Furthermore, it can inherit previous condensation results as clustering centroids when the condensed graph expands, thereby attaining an evolving capability. This methodology is supported by robust theoretical foundations and demonstrates superior empirical performance. Comprehensive experiments including real world scenario show that GECC achieves better performance than most state-of-the-art graph condensation methods while delivering an around 1000$\times$ speedup on large datasets.
Authors: Jonathan Shaki, Emanuele La Malfa, Michael Wooldridge, Sarit Kraus
Abstract: Binary relations, such as equality, are basic mathematical concepts that appear, implicitly or explicitly, in most benchmarks for Large Language Models (LLM). A recent trend in the literature is benchmarking LLMs on out-of-context learning, where the data is not presented in the prompt, but only during the model's training. However, existing works mostly focus on higher-order tasks, making it hard to interpret success or failure. In this work, we study how well can LLMs reason out-of-context on binary relations by only learning the representations of newly introduced tokens. Our experiments focus on equality ($=$), inequality ($<$), and inclusion ($\subset$) and the properties they satisfy, such as reflexivity, symmetry, transitivity, and logical complexity (e.g., the number of reasoning "hops"). We show that LLMs achieve better than random accuracy, but are still far from perfect, even on relatively simple reasoning tasks involving binary relations. We analyse the learned representations and show that LLMs encode useful information directly, arranging the embeddings according to the task.
Authors: Ho Kei Cheng, Alexander Schwing
Abstract: Minibatch optimal transport coupling straightens paths in unconditional flow matching. This leads to computationally less demanding inference as fewer integration steps and less complex numerical solvers can be employed when numerically solving an ordinary differential equation at test time. However, in the conditional setting, minibatch optimal transport falls short. This is because the default optimal transport mapping disregards conditions, resulting in a conditionally skewed prior distribution during training. In contrast, at test time, we have no access to the skewed prior, and instead sample from the full, unbiased prior distribution. This gap between training and testing leads to a subpar performance. To bridge this gap, we propose conditional optimal transport C^2OT that adds a conditional weighting term in the cost matrix when computing the optimal transport assignment. Experiments demonstrate that this simple fix works with both discrete and continuous conditions in 8gaussians-to-moons, CIFAR-10, ImageNet-32x32, and ImageNet-256x256. Our method performs better overall compared to the existing baselines across different function evaluation budgets. Code is available at https://hkchengrex.github.io/C2OT
Authors: Zhe Wang, Yanjun Qi
Abstract: Gradient optimization-based adversarial attack methods automate the learning of adversarial triggers to generate jailbreak prompts or leak system prompts. In this work, we take a closer look at the optimization objective of adversarial trigger learning and propose ATLA: Adversarial Trigger Learning with Augmented objectives. ATLA improves the negative log-likelihood loss used by previous studies into a weighted loss formulation that encourages the learned adversarial triggers to optimize more towards response format tokens. This enables ATLA to learn an adversarial trigger from just one query-response pair and the learned trigger generalizes well to other similar queries. We further design a variation to augment trigger optimization with an auxiliary loss that suppresses evasive responses. We showcase how to use ATLA to learn adversarial suffixes jailbreaking LLMs and to extract hidden system prompts. Empirically we demonstrate that ATLA consistently outperforms current state-of-the-art techniques, achieving nearly 100% success in attacking while requiring 80% fewer queries. ATLA learned jailbreak suffixes demonstrate high generalization to unseen queries and transfer well to new LLMs. We released our code \href{https://github.com/QData/ALTA_Augmented_Adversarial_Trigger_Learning}{here}.
URLs: https://github.com/QData/ALTA_Augmented_Adversarial_Trigger_Learning
Authors: Liya Guo, Zun Wang, Chang Liu, Junzhe Li, Pipi Hu, Yi Zhu
Abstract: The ensemble average of physical properties of molecules is closely related to the distribution of molecular conformations, and sampling such distributions is a fundamental challenge in physics and chemistry. Traditional methods like molecular dynamics (MD) simulations and Markov chain Monte Carlo (MCMC) sampling are commonly used but can be time-consuming and costly. Recently, diffusion models have emerged as efficient alternatives by learning the distribution of training data. Obtaining an unbiased target distribution is still an expensive task, primarily because it requires satisfying ergodicity. To tackle these challenges, we propose Potential Score Matching (PSM), an approach that utilizes the potential energy gradient to guide generative models. PSM does not require exact energy functions and can debias sample distributions even when trained on limited and biased data. Our method outperforms existing state-of-the-art (SOTA) models on the Lennard-Jones (LJ) potential, a commonly used toy model. Furthermore, we extend the evaluation of PSM to high-dimensional problems using the MD17 and MD22 datasets. The results demonstrate that molecular distributions generated by PSM more closely approximate the Boltzmann distribution compared to traditional diffusion models.
Authors: Gianluca Peri, Lorenzo Chicchi, Duccio Fanelli, Lorenzo Giambagli
Abstract: Architecture design and optimization are challenging problems in the field of artificial neural networks. Working in this context, we here present SPARCS (SPectral ARchiteCture Search), a novel architecture search protocol which exploits the spectral attributes of the inter-layer transfer matrices. SPARCS allows one to explore the space of possible architectures by spanning continuous and differentiable manifolds, thus enabling for gradient-based optimization algorithms to be eventually employed. With reference to simple benchmark models, we show that the newly proposed method yields a self-emerging architecture with a minimal degree of expressivity to handle the task under investigation and with a reduced parameter count as compared to other viable alternatives.
Authors: Deyuan Liu, Peng Sun, Xufeng Li, Tao Lin
Abstract: Diffusion models excel at generating high-dimensional data but fall short in training efficiency and representation quality compared to self-supervised methods. We identify a key bottleneck: the underutilization of high-quality, semantically rich representations during training notably slows down convergence. Our systematic analysis reveals a critical representation processing region -- primarily in the early layers -- where semantic and structural pattern learning takes place before generation can occur. To address this, we propose Embedded Representation Warmup (ERW), a plug-and-play framework where in the first stage we get the ERW module serves as a warmup that initializes the early layers of the diffusion model with high-quality, pretrained representations. This warmup minimizes the burden of learning representations from scratch, thereby accelerating convergence and boosting performance. Our theoretical analysis demonstrates that ERW's efficacy depends on its precise integration into specific neural network layers -- termed the representation processing region -- where the model primarily processes and transforms feature representations for later generation. We further establish that ERW not only accelerates training convergence but also enhances representation quality: empirically, our method achieves a 40$\times$ acceleration in training speed compared to REPA, the current state-of-the-art methods. Code is available at https://github.com/LINs-lab/ERW.
Authors: Evan Casey, Tianyu Zhang, Shu Ishida, William P. McCarthy, John Roger Thompson, Amir Khasahmadi, Joseph George Lambourne, Pradeep Kumar Jayaraman, Karl D. D. Willis
Abstract: We adapt alignment techniques from reasoning LLMs to the task of generating engineering sketch constraints found in computer-aided design (CAD) models. Engineering sketches consist of geometric primitives (e.g. points, lines) connected by constraints (e.g. perpendicular, tangent) that define the relationships between them. For a design to be easily editable, the constraints must effectively capture design intent, ensuring the geometry updates predictably when parameters change. Although current approaches can generate CAD designs, an open challenge remains to align model outputs with design intent, we label this problem 'design alignment'. A critical first step towards aligning generative CAD models is to generate constraints which fully-constrain all geometric primitives, without over-constraining or distorting sketch geometry. Using alignment techniques to train an existing constraint generation model with feedback from a constraint solver, we are able to fully-constrain 93% of sketches compared to 34% when using a naive supervised fine-tuning (SFT) baseline and only 8.9% without SFT. Our approach can be applied to any existing constraint generation model and sets the stage for further research bridging alignment strategies between the language and design domains. Additional results can be found at https://autodeskailab.github.io/aligning-constraint-generation/.
URLs: https://autodeskailab.github.io/aligning-constraint-generation/.
Authors: Lawrence Liu, Inesh Chakrabarti, Yixiao Li, Mengdi Wang, Tuo Zhao, Lin F. Yang
Abstract: Large language models (LLMs) exhibit remarkable performance across various natural language processing tasks but suffer from immense computational and memory demands, limiting their deployment in resource-constrained environments. To address this challenge, we propose NoWag: (Normalized Weight and Activation Guided Compression), a unified framework for zero-shot shape preserving compression algorithms. We compressed Llama-2 7B/13B/70B and Llama-3 8/70BB models, using two popular forms of shape-preserving compression, vector quantization NoWag-VQ (NoWag for Vector Quantization), and unstructured/semi-structured pruning NoWag-P (NoWag for Pruning). We found that NoWag-VQ significantly outperforms state-of-the-art zero shot VQ, and that NoWag-P performs competitively against state-of-the-art methods. These results suggest commonalities between these compression paradigms that could inspire future work. Our code is available at https://github.com/LawrenceRLiu/NoWag
Authors: Ching-Yi Lin, Sahil Shah
Abstract: Pre-trained vision transformers have achieved remarkable performance across various visual tasks but suffer from expensive computational and memory costs. While model quantization reduces memory usage by lowering precision, these models still incur significant computational overhead due to the dequantization before matrix operations. In this work, we analyze the computation graph and propose an integerization process based on operation reordering. Specifically, the process delays dequantization until after matrix operations. This enables integerized matrix multiplication and linear module by directly processing the quantized input. To validate our approach, we synthesize the self-attention module of ViT on a systolic array-based hardware. Experimental results show that our low-bit inference reduces per-PE power consumption for linear layer and matrix multiplication, bridging the gap between quantized models and efficient inference.
Authors: Chethan Krishnamurthy Ramanaik, Arjun Roy, Tobias Callies, Eirini Ntoutsi
Abstract: Adversarial robustness of deep autoencoders (AEs) remains relatively unexplored, even though their non-invertible nature poses distinct challenges. Existing attack algorithms during the optimization of imperceptible, norm-bounded adversarial perturbations to maximize output damage in AEs, often stop at sub-optimal attacks. We observe that the adversarial loss gradient vanishes when backpropagated through ill-conditioned layers. This issue arises from near-zero singular values in the Jacobians of these layers, which weaken the gradient signal during optimization. We introduce GRILL, a technique that locally restores gradient signals in ill-conditioned layers, enabling more effective norm-bounded attacks. Through extensive experiments on different architectures of popular AEs, under both sample-specific and universal attack setups, and across standard and adaptive attack settings, we show that our method significantly increases the effectiveness of our adversarial attacks, enabling a more rigorous evaluation of AE robustness.
Authors: Alexander Humer, Lukas Grasboeck, Ayech Benjeddou
Abstract: Today, machine learning is ubiquitous, and structural health monitoring (SHM) is no exception. Specifically, we address the problem of impact localization on shell-like structures, where knowledge of impact locations aids in assessing structural integrity. Impacts on thin-walled structures excite Lamb waves, which can be measured with piezoelectric sensors. Their dispersive characteristics make it difficult to detect and localize impacts by conventional methods. In the present contribution, we explore the localization of impacts using neural networks. In particular, we propose to use recurrent neural networks (RNNs) to estimate impact positions end-to-end, i.e., directly from sequential sensor data. We deal with comparatively long sequences of thousands of samples, since high sampling rate are needed to accurately capture elastic waves. For this reason, the proposed approach builds upon Gated Recurrent Units (GRUs), which are less prone to vanishing gradients as compared to conventional RNNs. Quality and quantity of data are crucial when training neural networks. Often, synthetic data is used, which inevitably introduces a reality gap. Here, by contrast, we train our networks using physical data from experiments, which requires automation to handle the large number of experiments needed. For this purpose, a robot is used to drop steel balls onto an aluminum plate equipped with piezoceramic sensors. Our results show remarkable accuracy in estimating impact positions, even with a comparatively small dataset.
Authors: Leon G\"otz, Marcel Kollovieh, Stephan G\"unnemann, Leo Schwinn
Abstract: Existing time series tokenization methods predominantly encode a constant number of samples into individual tokens. This inflexible approach can generate excessive tokens for even simple patterns like extended constant values, resulting in substantial computational overhead. Inspired by the success of byte pair encoding, we propose the first pattern-centric tokenization scheme for time series analysis. Based on a discrete vocabulary of frequent motifs, our method merges samples with underlying patterns into tokens, compressing time series adaptively. Exploiting our finite set of motifs and the continuous properties of time series, we further introduce conditional decoding as a lightweight yet powerful post-hoc optimization method, which requires no gradient computation and adds no computational overhead. On recent time series foundation models, our motif-based tokenization improves forecasting performance by 36% and boosts efficiency by 1990% on average. Conditional decoding further reduces MSE by up to 44%. In an extensive analysis, we demonstrate the adaptiveness of our tokenization to diverse temporal patterns, its generalization to unseen data, and its meaningful token representations capturing distinct time series properties, including statistical moments and trends.
Authors: Bo-Han Lai, Pin-Han Huang, Bo-Han Kung, Shang-Tse Chen
Abstract: Lipschitz neural networks are well-known for providing certified robustness in deep learning. In this paper, we present a novel, efficient Block Reflector Orthogonal (BRO) layer that enhances the capability of orthogonal layers on constructing more expressive Lipschitz neural architectures. In addition, by theoretically analyzing the nature of Lipschitz neural networks, we introduce a new loss function that employs an annealing mechanism to increase margin for most data points. This enables Lipschitz models to provide better certified robustness. By employing our BRO layer and loss function, we design BRONet - a simple yet effective Lipschitz neural network that achieves state-of-the-art certified robustness. Extensive experiments and empirical analysis on CIFAR-10/100, Tiny-ImageNet, and ImageNet validate that our method outperforms existing baselines. The implementation is available at https://github.com/ntuaislab/BRONet.
Authors: Yutong Chen, Jiandong Gao, Ji Wu
Abstract: R1-style Reinforcement Learning (RL) significantly enhances Large Language Models' reasoning capabilities, yet the mechanism behind rule-based RL remains unclear. We found that small-scale SFT has substantial influence on RL but shows poor efficiency. To explain our observations, we propose an analytical framework and compare the efficiency of SFT and RL by measuring \textbf{sample effect}. Our hypothetical analysis shows the potential to improve SFT efficiency. Guided by our analysis, we propose \textbf{Re-distillation}, a technique that aims to boost the effectiveness of small-scale distillation by sampling from the RL-trained policy. Re-distillation shows consistent surprising efficiency on three datasets and both Qwen\&Llama models: Re-distilled models matched RL performance with far fewer samples and less computation. As a result, on K\&K dataset, our re-distilled Qwen-2.5-1.5B model surpasses DeepSeek-V3-0324 with only 1K SFT samples. We demonstrate that re-distillation can be used to efficiently balance multiple goals in RL. Our work explains several interesting phenomena in R1-style RL, shedding light on the mechanisms behind its empirical success. Code is available at: https://github.com/on1262/deep-reasoning.
Authors: Afrah Farea, Saiful Khan, Reza Daryani, Emre Cenk Ersan, Mustafa Serdar Celebi
Abstract: We introduce neural network architectures that combine physics-informed neural networks (PINNs) with the immersed boundary method (IBM) to solve fluid-structure interaction (FSI) problems. Our approach features two distinct architectures: a Single-FSI network with a unified parameter space, and an innovative Eulerian-Lagrangian network that maintains separate parameter spaces for fluid and structure domains. We study each architecture using standard Tanh and adaptive B-spline activation functions. Empirical studies on a 2D cavity flow problem involving a moving solid structure show that the Eulerian-Lagrangian architecture performs significantly better. The adaptive B-spline activation further enhances accuracy by providing locality-aware representation near boundaries. While our methodology shows promising results in predicting the velocity field, pressure recovery remains challenging due to the absence of explicit force-coupling constraints in the current formulation. Our findings underscore the importance of domain-specific architectural design and adaptive activation functions for modeling FSI problems within the PINN framework.
Authors: Yuan Gao, Ruiqi Shu, Hao Wu, Fan Xu, Yanfei Xiang, Ruijian Gou, Qingsong Wen, Xian Wu, Kun Wang, Xiaomeng Huang
Abstract: Long-term, high-fidelity simulation of slow-changing physical systems, such as the ocean and climate, presents a fundamental challenge in scientific computing. Traditional autoregressive machine learning models often fail in these tasks as minor errors accumulate and lead to rapid forecast degradation. To address this problem, we propose NeuralOM, a general neural operator framework designed for simulating complex, slow-changing dynamics. NeuralOM's core consists of two key innovations: (1) a Progressive Residual Correction Framework that decomposes the forecasting task into a series of fine-grained refinement steps, effectively suppressing long-term error accumulation; and (2) a Physics-Guided Graph Network whose built-in adaptive messaging mechanism explicitly models multi-scale physical interactions, such as gradient-driven flows and multiplicative couplings, thereby enhancing physical consistency while maintaining computational efficiency. We validate NeuralOM on the challenging task of global Subseasonal-to-Seasonal (S2S) ocean simulation. Extensive experiments demonstrate that NeuralOM not only surpasses state-of-the-art models in forecast accuracy and long-term stability, but also excels in simulating extreme events. For instance, at a 60-day lead time, NeuralOM achieves a 13.3% lower RMSE compared to the best-performing baseline, offering a stable, efficient, and physically-aware paradigm for data-driven scientific computing. Code link: https://github.com/YuanGao-YG/NeuralOM.
Authors: Zeng Wang, Minghao Shao, Rupesh Karn, Likhitha Mankali, Jitendra Bhandari, Ramesh Karri, Ozgur Sinanoglu, Muhammad Shafique, Johann Knechtel
Abstract: Large Language Models (LLMs) offer transformative capabilities for hardware design automation, particularly in Verilog code generation. However, they also pose significant data security challenges, including Verilog evaluation data contamination, intellectual property (IP) design leakage, and the risk of malicious Verilog generation. We introduce SALAD, a comprehensive assessment that leverages machine unlearning to mitigate these threats. Our approach enables the selective removal of contaminated benchmarks, sensitive IP and design artifacts, or malicious code patterns from pre-trained LLMs, all without requiring full retraining. Through detailed case studies, we demonstrate how machine unlearning techniques effectively reduce data security risks in LLM-aided hardware design.
Authors: Fang Wang, Paolo Ceravolo, Ernesto Damiani
Abstract: Predictive Business Process Monitoring (PBPM) aims to forecast future outcomes of ongoing business processes. However, existing methods often lack flexibility to handle real-world challenges such as simultaneous events, class imbalance, and multi-level attributes. While prior work has explored static encoding schemes and fixed LSTM architectures, they struggle to support adaptive representations and generalize across heterogeneous datasets. To address these limitations, we propose a suite of dynamic LSTM HyperModels that integrate two-level hierarchical encoding for event and sequence attributes, character-based decomposition of event labels, and novel pseudo-embedding techniques for durations and attribute correlations. We further introduce specialized LSTM variants for simultaneous event modeling, leveraging multidimensional embeddings and time-difference flag augmentation. Experimental validation on four public and real-world datasets demonstrates up to 100% accuracy on balanced datasets and F1 scores exceeding 86\% on imbalanced ones. Our approach advances PBPM by offering modular and interpretable models better suited for deployment in complex settings. Beyond PBPM, it contributes to the broader AI community by improving temporal outcome prediction, supporting data heterogeneity, and promoting explainable process intelligence frameworks.
Authors: Seyedhamidreza Mousavi, Seyedali Mousavi, Masoud Daneshtalab
Abstract: Adversarial Robustness Distillation (ARD) has emerged as an effective method to enhance the robustness of lightweight deep neural networks against adversarial attacks. Current ARD approaches have leveraged a large robust teacher network to train one robust lightweight student. However, due to the diverse range of edge devices and resource constraints, current approaches require training a new student network from scratch to meet specific constraints, leading to substantial computational costs and increased CO2 emissions. This paper proposes Progressive Adversarial Robustness Distillation (ProARD), enabling the efficient one-time training of a dynamic network that supports a diverse range of accurate and robust student networks without requiring retraining. We first make a dynamic deep neural network based on dynamic layers by encompassing variations in width, depth, and expansion in each design stage to support a wide range of architectures. Then, we consider the student network with the largest size as the dynamic teacher network. ProARD trains this dynamic network using a weight-sharing mechanism to jointly optimize the dynamic teacher network and its internal student networks. However, due to the high computational cost of calculating exact gradients for all the students within the dynamic network, a sampling mechanism is required to select a subset of students. We show that random student sampling in each iteration fails to produce accurate and robust students.
Authors: Kangkang Lu, Yanhua Yu, Zhiyong Huang, Tat-Seng Chua
Abstract: Spectral Graph Neural Networks (SGNNs) have achieved remarkable performance in tasks such as node classification due to their ability to learn flexible filters. Typically, these filters are learned under the supervision of downstream tasks, enabling SGNNs to adapt to diverse structural patterns. However, in scenarios with limited labeled data, SGNNs often struggle to capture the optimal filter shapes, resulting in degraded performance, especially on graphs with heterophily. Meanwhile, the rapid progress of Large Language Models (LLMs) has opened new possibilities for enhancing graph learning without modifying graph structure or requiring task-specific training. In this work, we propose a novel framework that leverages LLMs to estimate the homophily level of a graph and uses this global structural prior to guide the construction of spectral filters. Specifically, we design a lightweight and plug-and-play pipeline where a small set of labeled node pairs is formatted as natural language prompts for the LLM, which then predicts the graph's homophily ratio. This estimated value informs the spectral filter basis, enabling SGNNs to adapt more effectively to both homophilic and heterophilic structures. Extensive experiments on multiple benchmark datasets demonstrate that our LLM-assisted spectral framework consistently improves performance over strong SGNN baselines. Importantly, this enhancement incurs negligible computational and monetary cost, making it a practical solution for real-world graph applications.
Authors: Sunil Kumar, Bowen Zhao, Leo Dirac, Paulina Varshavskaya
Abstract: Despite tremendous recent advances in large model reasoning ability, vision-language models (VLMs) still struggle with detailed visual reasoning, especially when compute resources are limited. To address this challenge, we draw inspiration from methods like Deepseek-r1 for VLMs and train smaller-scale models with Group Relative Policy Optimization (GRPO) to use external tools such as zoom. The greatest benefit is obtained with a combination of GRPO learning, a simple reward structure, a simplified tool-calling interface, allocating additional tokens to the result of the tool call, and a training data mix that over-represents visually difficult examples. Compared to similarly-sized baseline models, our method achieves better performance on some visual question-answering (VQA) tasks, thanks to the detailed visual information gathered from the external tool.
Authors: Zihan Tan, Suyuan Huang, Guancheng Wan, Wenke Huang, He Li, Mang Ye
Abstract: Federated Graph Learning (FGL) combines the privacy-preserving capabilities of federated learning (FL) with the strong graph modeling capability of Graph Neural Networks (GNNs). Current research addresses subgraph-FL from the structural perspective, neglecting the propagation of graph signals on spatial and spectral domains of the structure. From a spatial perspective, subgraph-FL introduces edge disconnections between clients, leading to disruptions in label signals and a degradation in the semantic knowledge of the global GNN. From a spectral perspective, spectral heterogeneity causes inconsistencies in signal frequencies across subgraphs, which makes local GNNs overfit the local signal propagation schemes. As a result, spectral client drift occurs, undermining global generalizability. To tackle the challenges, we propose a global knowledge repository to mitigate the challenge of poor semantic knowledge caused by label signal disruption. Furthermore, we design a frequency alignment to address spectral client drift. The combination of Spatial and Spectral strategies forms our framework S2FGL. Extensive experiments on multiple datasets demonstrate the superiority of S2FGL. The code is available at https://github.com/Wonder7racer/S2FGL.git.
Authors: Shengjian Chen
Abstract: We point out that neural networks are not black boxes, and their generalization stems from the ability to dynamically map a dataset to the extrema of the model function. We further prove that the number of extrema in a neural network is positively correlated with the number of its parameters. We then propose a new algorithm that is significantly different from back-propagation algorithm, which mainly obtains the values of parameters by solving a system of linear equations. Some difficult situations, such as gradient vanishing and overfitting, can be reasonably explained and dealt with in this framework.
Authors: Janna Lu
Abstract: Large language models (LLMs) have demonstrated remarkable capabilities across diverse tasks, but their ability to forecast future events remains understudied. A year ago, large language models struggle to come close to the accuracy of a human crowd. I evaluate state-of-the-art LLMs on 464 forecasting questions from Metaculus, comparing their performance against top forecasters. Frontier models achieve Brier scores that ostensibly surpass the human crowd but still significantly underperform a group of experts.
Authors: Siddhant Deshpande, Yalemzerf Getnet, Waltenegus Dargie
Abstract: With the proliferation of wireless electrocardiogram (ECG) systems for health monitoring and authentication, protecting signal integrity against tampering is becoming increasingly important. This paper analyzes the performance of CNN, ResNet, and hybrid Transformer-CNN models for tamper detection. It also evaluates the performance of a Siamese network for ECG based identity verification. Six tampering strategies, including structured segment substitutions and random insertions, are emulated to mimic real world attacks. The one-dimensional ECG signals are transformed into a two dimensional representation in the time frequency domain using the continuous wavelet transform (CWT). The models are trained and evaluated using ECG data from 54 subjects recorded in four sessions 2019 to 2025 outside of clinical settings while the subjects performed seven different daily activities. Experimental results show that in highly fragmented manipulation scenarios, CNN, FeatCNN-TranCNN, FeatCNN-Tran and ResNet models achieved an accuracy exceeding 99.5 percent . Similarly, for subtle manipulations (for example, 50 percent from A and 50 percent from B and, 75 percent from A and 25 percent from B substitutions) our FeatCNN-TranCNN model demonstrated consistently reliable performance, achieving an average accuracy of 98 percent . For identity verification, the pure Transformer-Siamese network achieved an average accuracy of 98.30 percent . In contrast, the hybrid CNN-Transformer Siamese model delivered perfect verification performance with 100 percent accuracy.
Authors: Wenxuan Zhou, Shujian Zhang, Brice Magdalou, John Lambert, Ehsan Amid, Richard Nock, Andrew Hard
Abstract: In this paper, we show that direct preference optimization (DPO) is a very specific form of a connection between two major theories in the ML context of learning from preferences: loss functions (Savage) and stochastic choice (Doignon-Falmagne and Machina). The connection is established for all of Savage's losses and at this level of generality, (i) it includes support for abstention on the choice theory side, (ii) it includes support for non-convex objectives on the ML side, and (iii) it allows to frame for free some notable extensions of the DPO setting, including margins and corrections for length. Getting to understand how DPO operates from a general principled perspective is crucial because of the huge and diverse application landscape of models, because of the current momentum around DPO, but also -- and importantly -- because many state of the art variations on DPO definitely occupy a small region of the map that we cover. It also helps to understand the pitfalls of departing from this map, and figure out workarounds.
Authors: Junaid Iqbal Khan
Abstract: Approximate machine unlearning (AMU) enables models to `forget' specific training data through specialized fine-tuning on a retained (and forget) subset of training set. However, processing this large retained subset still dominates computational runtime, while reductions of unlearning epochs also remain a challenge. In this paper, we propose two complementary methods to accelerate arbitrary classification-oriented AMU method. First, \textbf{Blend}, a novel distribution-matching dataset condensation (DC), merges visually similar images with shared blend-weights to significantly reduce the retained set size. It operates with minimal pre-processing overhead and is orders of magnitude faster than state-of-the-art DC methods. Second, our loss-centric method, \textbf{Accelerated-AMU (A-AMU)}, augments the AMU objective to quicken convergence. A-AMU achieves this by combining a steepened primary loss to expedite forgetting with a differentiable regularizer that matches the loss distributions of forgotten and in-distribution unseen data. Our extensive experiments demonstrate that this dual approach of data and loss-centric optimization dramatically reduces end-to-end unlearning latency across both single and multi-round scenarios, all while preserving model utility and privacy. To our knowledge, this is the first work to systematically tackle unlearning efficiency by jointly designing a specialized dataset condensation technique with a dedicated accelerated loss function. Code is available at https://github.com/algebraicdianuj/DC_Unlearning.
Authors: Jiamin Wu, Zichen Ren, Junyu Wang, Pengyu Zhu, Yonghao Song, Mianxin Liu, Qihao Zheng, Lei Bai, Wanli Ouyang, Chunfeng Song
Abstract: Non-invasive Brain-Computer Interfaces (BCI) offer a safe and accessible means of connecting the human brain to external devices, with broad applications in home and clinical settings to enhance human capabilities. However, the high noise level and limited task-specific data in non-invasive signals constrain decoding capabilities. Recently, the adoption of self-supervised pre-training is transforming the landscape of non-invasive BCI research, enabling the development of brain foundation models to capture generic neural representations from large-scale unlabeled electroencephalography (EEG) signals with substantial noises. However, despite these advances, the field currently lacks comprehensive, practical and extensible benchmarks to assess the utility of the public foundation models across diverse BCI tasks, hindering their widespread adoption. To address this challenge, we present AdaBrain-Bench, a large-scale standardized benchmark to systematically evaluate brain foundation models in widespread non-invasive BCI tasks. AdaBrain-Bench encompasses a diverse collection of representative BCI decoding datasets spanning 7 key applications. It introduces a streamlined task adaptation pipeline integrated with multi-dimensional evaluation metrics and a set of adaptation tools. The benchmark delivers an inclusive framework for assessing generalizability of brain foundation models across key transfer settings, including cross-subject, multi-subject, and few-shot scenarios. We leverage AdaBrain-Bench to evaluate a suite of publicly available brain foundation models and offer insights into practices for selecting appropriate models in various scenarios. We make our benchmark pipeline available to enable reproducible research and external use, offering a continuously evolving platform to foster progress toward robust and generalized neural decoding solutions.
Authors: Mingqi Wu, Zhihao Zhang, Qiaole Dong, Zhiheng Xi, Jun Zhao, Senjie Jin, Xiaoran Fan, Yuhao Zhou, Huijie Lv, Ming Zhang, Yanwei Fu, Qin Liu, Songyang Zhang, Qi Zhang
Abstract: Reasoning in large language models has long been a central research focus, and recent studies employing reinforcement learning (RL) have introduced diverse methods that yield substantial performance gains with minimal or even no external supervision. Surprisingly, some studies even suggest that random or incorrect reward signals can enhance performance. However, these breakthroughs are predominantly observed for the mathematically strong Qwen2.5 series on benchmarks such as MATH-500, AMC, and AIME, and seldom transfer to models like Llama, which warrants a more in-depth investigation. In this work, our empirical analysis reveals that pre-training on massive web-scale corpora leaves Qwen2.5 susceptible to data contamination in widely used benchmarks. Consequently, conclusions derived from contaminated benchmarks on Qwen2.5 series may be unreliable. To obtain trustworthy evaluation results, we introduce a generator that creates fully clean arithmetic problems of arbitrary length and difficulty, dubbed RandomCalculation. Using this leakage-free dataset, we show that only accurate reward signals yield steady improvements that surpass the base model's performance boundary in mathematical reasoning, whereas random or incorrect rewards do not. Moreover, we conduct more fine-grained analyses to elucidate the factors underlying the different performance observed on the MATH-500 and RandomCalculation benchmarks. Consequently, we recommend that future studies evaluate models on uncontaminated benchmarks and, when feasible, test various model series to ensure trustworthy conclusions about RL and related methods.
Authors: Yaowen Hu, Wenxuan Tu, Yue Liu, Miaomiao Li, Wenpeng Lu, Zhigang Luo, Xinwang Liu, Ping Chen
Abstract: Deep graph clustering (DGC) for attribute-missing graphs is an unsupervised task aimed at partitioning nodes with incomplete attributes into distinct clusters. Addressing this challenging issue is vital for practical applications. However, research in this area remains underexplored. Existing imputation methods for attribute-missing graphs often fail to account for the varying amounts of information available across node neighborhoods, leading to unreliable results, especially for nodes with insufficient known neighborhood. To address this issue, we propose a novel method named Divide-Then-Rule Graph Completion (DTRGC). This method first addresses nodes with sufficient known neighborhood information and treats the imputed results as new knowledge to iteratively impute more challenging nodes, while leveraging clustering information to correct imputation errors. Specifically, Dynamic Cluster-Aware Feature Propagation (DCFP) initializes missing node attributes by adjusting propagation weights based on the clustering structure. Subsequently, Hierarchical Neighborhood-aware Imputation (HNAI) categorizes attribute-missing nodes into three groups based on the completeness of their neighborhood attributes. The imputation is performed hierarchically, prioritizing the groups with nodes that have the most available neighborhood information. The cluster structure is then used to refine the imputation and correct potential errors. Finally, Hop-wise Representation Enhancement (HRE) integrates information across multiple hops, thereby enriching the expressiveness of node representations. Experimental results on six widely used graph datasets show that DTRGC significantly improves the clustering performance of various DGC methods under attribute-missing graphs.
Authors: Wei Fan, JinYi Yoon, Xiaochang Li, Huajie Shao, Bo Ji
Abstract: Split Learning (SL) is an emerging privacy-preserving machine learning technique that enables resource constrained edge devices to participate in model training by partitioning a model into client-side and server-side sub-models. While SL reduces computational overhead on edge devices, it encounters significant challenges in heterogeneous environments where devices vary in computing resources, communication capabilities, environmental conditions, and privacy requirements. Although recent studies have explored heterogeneous SL frameworks that optimize split points for devices with varying resource constraints, they often neglect personalized privacy requirements and local model customization under varying environmental conditions. To address these limitations, we propose P3SL, a Personalized Privacy-Preserving Split Learning framework designed for heterogeneous, resource-constrained edge device systems. The key contributions of this work are twofold. First, we design a personalized sequential split learning pipeline that allows each client to achieve customized privacy protection and maintain personalized local models tailored to their computational resources, environmental conditions, and privacy needs. Second, we adopt a bi-level optimization technique that empowers clients to determine their own optimal personalized split points without sharing private sensitive information (i.e., computational resources, environmental conditions, privacy requirements) with the server. This approach balances energy consumption and privacy leakage risks while maintaining high model accuracy. We implement and evaluate P3SL on a testbed consisting of 7 devices including 4 Jetson Nano P3450 devices, 2 Raspberry Pis, and 1 laptop, using diverse model architectures and datasets under varying environmental conditions.
Authors: Zhuokun Chen, Zeren Chen, Jiahao He, Mingkui Tan, Jianfei Cai, Bohan Zhuang
Abstract: Chain-of-thought (CoT) reasoning enhances the problem-solving capabilities of large language models by encouraging step-by-step intermediate reasoning during inference. While effective, CoT introduces substantial computational overhead due to its reliance on autoregressive decoding over long token sequences. Existing acceleration strategies either reduce sequence length through early stopping or compressive reward designs, or improve decoding speed via speculative decoding with smaller models. However, speculative decoding suffers from limited speedup when the agreement between small and large models is low, and fails to exploit the potential advantages of small models in producing concise intermediate reasoning. In this paper, we present R-Stitch, a token-level, confidence-based hybrid decoding framework that accelerates CoT inference by switching between a small language model (SLM) and a large language model (LLM) along the reasoning trajectory. R-Stitch uses the SLM to generate tokens by default and delegates to the LLM only when the SLM's confidence falls below a threshold. This design avoids full-sequence rollback and selectively invokes the LLM on uncertain steps, preserving both efficiency and answer quality. R-Stitch is model-agnostic, training-free, and compatible with standard decoding pipelines. Experiments on math reasoning benchmarks demonstrate that R-Stitch achieves up to 85\% reduction in inference latency with negligible accuracy drop, highlighting its practical effectiveness in accelerating CoT reasoning.
Authors: Melih Barsbey, Lucas Prieto, Stefanos Zafeiriou, Tolga Birdal
Abstract: Robustness and resource-efficiency are two highly desirable properties for modern machine learning models. However, achieving them jointly remains a challenge. In this paper, we identify high learning rates as a facilitator for simultaneously achieving robustness to spurious correlations and network compressibility. We demonstrate that large learning rates also produce desirable representation properties such as invariant feature utilization, class separation, and activation sparsity. Our findings indicate that large learning rates compare favorably to other hyperparameters and regularization methods, in consistently satisfying these properties in tandem. In addition to demonstrating the positive effect of large learning rates across diverse spurious correlation datasets, models, and optimizers, we also present strong evidence that the previously documented success of large learning rates in standard classification tasks is related to addressing hidden/rare spurious correlations in the training dataset. Our investigation of the mechanisms underlying this phenomenon reveals the importance of confident mispredictions of bias-conflicting samples under large learning rates.
Authors: Zhongzheng Yuan, Lianshuai Guo, Xunkai Li, Yinlin Zhu, Wenyu Wang, Meixia Qu
Abstract: Federated Graph Learning (FGL) is a distributed learning paradigm that enables collaborative training over large-scale subgraphs located on multiple local systems. However, most existing FGL approaches rely on synchronous communication, which leads to inefficiencies and is often impractical in real-world deployments. Meanwhile, current asynchronous federated learning (AFL) methods are primarily designed for conventional tasks such as image classification and natural language processing, without accounting for the unique topological properties of graph data. Directly applying these methods to graph learning can possibly result in semantic drift and representational inconsistency in the global model. To address these challenges, we propose FedSA-GCL, a semi-asynchronous federated framework that leverages both inter-client label distribution divergence and graph topological characteristics through a novel ClusterCast mechanism for efficient training. We evaluate FedSA-GCL on multiple real-world graph datasets using the Louvain and Metis split algorithms, and compare it against 9 baselines. Extensive experiments demonstrate that our method achieves strong robustness and outstanding efficiency, outperforming the baselines by an average of 2.92% with the Louvain and by 3.4% with the Metis.
Authors: Xing Hu, Haodong Chen, Qianqian Duan, Choon Ki Ahn, Huiliang Shang, Dawei Zhang Zhang
Abstract: With the global population increasing and arable land resources becoming increasingly limited, smart and precision agriculture have emerged as essential directions for sustainable agricultural development. Artificial intelligence (AI), particularly deep learning models, has been widely adopted in applications such as crop monitoring, pest detection, and yield prediction. Among recent generative models, diffusion models have demonstrated considerable potential in agricultural image processing, data augmentation, and remote sensing analysis. Compared to traditional generative adversarial networks (GANs), diffusion models exhibit greater training stability and superior image generation quality, effectively addressing challenges such as limited annotated datasets and imbalanced sample distributions in agricultural scenarios. This paper reviews recent advancements in the application of diffusion models within agriculture, focusing on their roles in crop disease and pest detection, remote sensing image enhancement, crop growth prediction, and agricultural resource management. Empirical studies show that diffusion models significantly enhance the performance of downstream models by improving accuracy, robustness, and generalization in tasks involving image synthesis, augmentation, and denoising under complex environmental conditions. Despite ongoing challenges in computational efficiency and domain generalization, diffusion models are expected to play an increasingly important role in the future of intelligent agriculture. As the technology continues to evolve, it holds substantial promise for addressing pressing global issues in food security and environmental sustainability.
Authors: Rongyao Cai, Ming Jin, Qingsong Wen, Kexin Zhang
Abstract: Domain shift poses a fundamental challenge in time series analysis, where models trained on source domain often fail dramatically when applied in target domain with different yet similar distributions. While current unsupervised domain adaptation (UDA) methods attempt to align cross-domain feature distributions, they typically treat features as indivisible entities, ignoring their intrinsic compositions that govern domain adaptation. We introduce DARSD, a novel UDA framework with theoretical explainability that explicitly realizes UDA tasks from the perspective of representation space decomposition. Our core insight is that effective domain adaptation requires not just alignment, but principled disentanglement of transferable knowledge from mixed representations. DARSD consists of three synergistic components: (I) An adversarial learnable common invariant basis that projects original features into a domain-invariant subspace while preserving semantic content; (II) A prototypical pseudo-labeling mechanism that dynamically separates target features based on confidence, hindering error accumulation; (III) A hybrid contrastive optimization strategy that simultaneously enforces feature clustering and consistency while mitigating emerging distribution gaps. Comprehensive experiments conducted on four benchmarks (WISDM, HAR, HHAR, and MFD) demonstrate DARSD's superiority against 12 UDA algorithms, achieving optimal performance in 35 out of 53 scenarios and ranking first across all benchmarks.
Authors: Fang Wang, Paolo Ceravolo, Ernesto Damiani
Abstract: We propose HGCN(O), a self-tuning toolkit using Graph Convolutional Network (GCN) models for event sequence prediction. Featuring four GCN architectures (O-GCN, T-GCN, TP-GCN, TE-GCN) across the GCNConv and GraphConv layers, our toolkit integrates multiple graph representations of event sequences with different choices of node- and graph-level attributes and in temporal dependencies via edge weights, optimising prediction accuracy and stability for balanced and unbalanced datasets. Extensive experiments show that GCNConv models excel on unbalanced data, while all models perform consistently on balanced data. Experiments also confirm the superior performance of HGCN(O) over traditional approaches. Applications include Predictive Business Process Monitoring (PBPM), which predicts future events or states of a business process based on event logs.
Authors: Judy X Yang
Abstract: Hyperspectral imaging offers detailed spectral information for mineral mapping; however, weak mineral signatures are often masked by noisy and redundant bands, limiting detection performance. To address this, we propose a two-stage integrated framework for enhanced mineral detection in the Cuprite mining district. In the first stage, we compute the signal-to-noise ratio (SNR) for each spectral band and apply a phase-locked thresholding technique to discard low-SNR bands, effectively removing redundancy and suppressing background noise. Savitzky-Golay filtering is then employed for spectral smoothing, serving a dual role first to stabilize trends during band selection, and second to preserve fine-grained spectral features during preprocessing. In the second stage, the refined HSI data is reintroduced into the model, where KMeans clustering is used to extract 12 endmember spectra (W1 custom), followed by non negative least squares (NNLS) for abundance unmixing. The resulting endmembers are quantitatively compared with laboratory spectra (W1 raw) using cosine similarity and RMSE metrics. Experimental results confirm that our proposed pipeline improves unmixing accuracy and enhances the detection of weak mineral zones. This two-pass strategy demonstrates a practical and reproducible solution for spectral dimensionality reduction and unmixing in geological HSI applications.
Authors: Ruichen Xu, Kexin Chen
Abstract: Modern large language models excel in knowledge-intensive tasks, yet how transformers acquire (store) knowledge during pre-training and extract (retrieve) it during post-fine-tuning inference remains theoretically opaque. While prior theoretical work has begun to investigate these questions through the analysis of training dynamics, such studies are limited to single-layer, attention-only architectures. However, most existing studies suggest that MLPs are the most contributing components for storing knowledge in transformer-based language models. Meanwhile, our empirical investigations reveal that such simplified models, when trained using standard next-token prediction objectives, may be incapable of acquiring or extracting factual knowledge. To overcome this limitation, we introduce a tractable one-layer transformer framework that crucially incorporates both self-attention and MLP modules. By tracking its gradient dynamics, we establish convergence and generalization guarantees that illuminate the ability of knowledge acquisition and extraction. We prove that 1) Transformers can achieve near-optimal training loss during pre-training, signifying effective knowledge acquisition; 2) With a large fine-tuning dataset and specific data multiplicity conditions met, transformers can achieve low generalization error when tested on factual knowledge learned during pre-training but not reinforced during the fine-tuning, indicating successful knowledge extraction; 3) When the conditions are not satisfied, transformers exhibit high generalization loss, resulting in hallucinations. Our analysis includes both full fine-tuning and low-rank fine-tuning. Furthermore, our analysis offers theoretical insights into several pertinent empirical phenomena, such as the role of learning rate schedules. Experiments on synthetic and real-world PopQA datasets with GPT-2 and Llama-3.2-1B validate our results.
Authors: Xuan Liu, Siru Ouyang, Xianrui Zhong, Jiawei Han, Huimin Zhao
Abstract: Large language models (LLMs) have gained significant attention in chemistry. However, most existing datasets center on molecular-level property prediction and overlook the role of fine-grained functional group (FG) information. Incorporating FG-level data can provide valuable prior knowledge that links molecular structures with textual descriptions, which can be used to build more interpretable, structure-aware LLMs for reasoning on molecule-related tasks. Moreover, LLMs can learn from such fine-grained information to uncover hidden relationships between specific functional groups and molecular properties, thereby advancing molecular design and drug discovery. Here, we introduce FGBench, a dataset comprising 625K molecular property reasoning problems with functional group information. Functional groups are precisely annotated and localized within the molecule, which ensures the dataset's interoperability thereby facilitating further multimodal applications. FGBench includes both regression and classification tasks on 245 different functional groups across three categories for molecular property reasoning: (1) single functional group impacts, (2) multiple functional group interactions, and (3) direct molecular comparisons. In the benchmark of state-of-the-art LLMs on 7K curated data, the results indicate that current LLMs struggle with FG-level property reasoning, highlighting the need to enhance reasoning capabilities in LLMs for chemistry tasks. We anticipate that the methodology employed in FGBench to construct datasets with functional group-level information will serve as a foundational framework for generating new question-answer pairs, enabling LLMs to better understand fine-grained molecular structure-property relationships. The dataset and evaluation code are available at https://github.com/xuanliugit/FGBench.
Authors: Zhuo Yang, Jiaqing Xie, Shuaike Shen, Daolang Wang, Yeyun Chen, Ben Gao, Shuzhou Sun, Biqing Qi, Dongzhan Zhou, Lei Bai, Linjiang Chen, Shufei Zhang, Jun Jiang, Tianfan Fu, Yuqiang Li
Abstract: Deep learning holds immense promise for spectroscopy, yet research and evaluation in this emerging field often lack standardized formulations. To address this issue, we introduce SpectrumLab, a pioneering unified platform designed to systematize and accelerate deep learning research in spectroscopy. SpectrumLab integrates three core components: a comprehensive Python library featuring essential data processing and evaluation tools, along with leaderboards; an innovative SpectrumAnnotator module that generates high-quality benchmarks from limited seed data; and SpectrumBench, a multi-layered benchmark suite covering 14 spectroscopic tasks and over 10 spectrum types, featuring spectra curated from over 1.2 million distinct chemical substances. Thorough empirical studies on SpectrumBench with 18 cutting-edge multimodal LLMs reveal critical limitations of current approaches. We hope SpectrumLab will serve as a crucial foundation for future advancements in deep learning-driven spectroscopy.
Authors: Xin Ding, Yun Chen, Yongwei Wang, Kao Zhang, Sen Zhang, Peibei Cao, Xiangxue Wang
Abstract: Recent advances in conditional generative modeling have introduced Continuous conditional Generative Adversarial Network (CcGAN) and Continuous Conditional Diffusion Model (CCDM) for estimating high-dimensional data distributions conditioned on scalar, continuous regression labels (e.g., angles, ages, or temperatures). However, these approaches face fundamental limitations: CcGAN suffers from data imbalance due to fixed-size vicinity constraints, while CCDM requires computationally expensive iterative sampling. We present CcGAN-AVAR, an enhanced CcGAN framework that addresses both challenges: (1) leveraging the GAN framework's native one-step generation to overcome CCDMs' sampling bottleneck (achieving 300x-2000x faster inference), while (2) two novel components specifically target data imbalance - an adaptive vicinity mechanism that dynamically adjusts vicinity's size, and a multi-task discriminator that constructs two regularization terms (through auxiliary regression and density ratio estimation) to significantly improve generator training. Extensive experiments on four benchmark datasets (64x64 to 192x192 resolution) across eight challenging imbalanced settings demonstrate that CcGAN-AVAR achieves state-of-the-art generation quality while maintaining sampling efficiency.
Authors: Alexander Norcliffe, Changhee Lee, Fergus Imrie, Mihaela van der Schaar, Pietro Lio
Abstract: Active Feature Acquisition is an instance-wise, sequential decision making problem. The aim is to dynamically select which feature to measure based on current observations, independently for each test instance. Common approaches either use Reinforcement Learning, which experiences training difficulties, or greedily maximize the conditional mutual information of the label and unobserved features, which makes myopic acquisitions. To address these shortcomings, we introduce a latent variable model, trained in a supervised manner. Acquisitions are made by reasoning about the features across many possible unobserved realizations in a stochastic latent space. Extensive evaluation on a large range of synthetic and real datasets demonstrates that our approach reliably outperforms a diverse set of baselines.
Authors: Parth Naik, Harikrishnan N B
Abstract: We propose a novel classification framework grounded in symbolic dynamics and data compression using chaotic maps. The core idea is to model each class by generating symbolic sequences from thresholded real-valued training data, which are then evolved through a one-dimensional chaotic map. For each class, we compute the transition probabilities of symbolic patterns (e.g., `00', `01', `10', and `11' for the second return map) and aggregate these statistics to form a class-specific probabilistic model. During testing phase, the test data are thresholded and symbolized, and then encoded using the class-wise symbolic statistics via back iteration, a dynamical reconstruction technique. The predicted label corresponds to the class yielding the shortest compressed representation, signifying the most efficient symbolic encoding under its respective chaotic model. This approach fuses concepts from dynamical systems, symbolic representations, and compression-based learning. We evaluate the proposed method: \emph{ChaosComp} on both synthetic and real-world datasets, demonstrating competitive performance compared to traditional machine learning algorithms (e.g., macro F1-scores for the proposed method on Breast Cancer Wisconsin = 0.9531, Seeds = 0.9475, Iris = 0.8469 etc.). Rather than aiming for state-of-the-art performance, the goal of this research is to reinterpret the classification problem through the lens of dynamical systems and compression, which are foundational perspectives in learning theory and information processing.
Authors: Kunyu Zhang, Lin Gu, Liangchen Liu, Yingke Chen, Binyang Wang, Jin Yan, Yingying Zhu
Abstract: Many previous studies have proposed extracting image labels from clinical notes to create large-scale medical image datasets at a low cost. However, these approaches inherently suffer from label noise due to uncertainty from the clinical experts. When radiologists and physicians analyze medical images to make diagnoses, they often include uncertainty-aware notes such as ``maybe'' or ``not excluded''. Unfortunately, current text-mining methods overlook these nuances, resulting in the creation of noisy labels. Existing methods for handling noisy labels in medical image analysis, which typically address the problem through post-processing techniques, have largely ignored the important issue of expert-driven uncertainty contributing to label noise. To better incorporate the expert-written uncertainty in clinical notes into medical image analysis and address the label noise issue, we first examine the impact of clinical expert uncertainty on label noise. We then propose a clinical expert uncertainty-aware benchmark, along with a label smoothing method, which significantly improves performance compared to current state-of-the-art approaches.
Authors: Prateek Mittal, Puneet Goyal, Joohi Chauhan
Abstract: This study introduces a novel multimodal food recognition framework that effectively combines visual and textual modalities to enhance classification accuracy and robustness. The proposed approach employs a dynamic multimodal fusion strategy that adaptively integrates features from unimodal visual inputs and complementary textual metadata. This fusion mechanism is designed to maximize the use of informative content, while mitigating the adverse impact of missing or inconsistent modality data. The framework was rigorously evaluated on the UPMC Food-101 dataset and achieved unimodal classification accuracies of 73.60% for images and 88.84% for text. When both modalities were fused, the model achieved an accuracy of 97.84%, outperforming several state-of-the-art methods. Extensive experimental analysis demonstrated the robustness, adaptability, and computational efficiency of the proposed settings, highlighting its practical applicability to real-world multimodal food-recognition scenarios.
Authors: Michael T. Wojnowicz, Kaitlin Gili, Preetish Rath, Eric Miller, Jeffrey Miller, Clifford Hancock, Meghan O'Donovan, Seth Elkin-Frankston, Tad T. Bruny\'e, Michael C. Hughes
Abstract: We seek a computationally efficient model for a collection of time series arising from multiple interacting entities (a.k.a. "agents"). Recent models of temporal patterns across individuals fail to incorporate explicit system-level collective behavior that can influence the trajectories of individual entities. To address this gap in the literature, we present a new hierarchical switching-state model that can be trained in an unsupervised fashion to simultaneously learn both system-level and individual-level dynamics. We employ a latent system-level discrete state Markov chain that provides top-down influence on latent entity-level chains which in turn govern the emission of each observed time series. Recurrent feedback from the observations to the latent chains at both entity and system levels allows recent situational context to inform how dynamics unfold at all levels in bottom-up fashion. We hypothesize that including both top-down and bottom-up influences on group dynamics will improve interpretability of the learned dynamics and reduce error when forecasting. Our hierarchical switching recurrent dynamical model can be learned via closed-form variational coordinate ascent updates to all latent chains that scale linearly in the number of entities. This is asymptotically no more costly than fitting a separate model for each entity. Analysis of both synthetic data and real basketball team movements suggests our lean parametric model can achieve competitive forecasts compared to larger neural network models that require far more computational resources. Further experiments on soldier data as well as a synthetic task with 64 cooperating entities show how our approach can yield interpretable insights about team dynamics over time.
Authors: Nemat Gholinejad, Mostafa Haghir Chehreghani
Abstract: In recent years, graph neural networks (GNNs) have become a popular tool to improve the accuracy and performance of recommender systems. Modern recommender systems are not only designed to serve end users, but also to benefit other participants, such as items and item providers. These participants may have different or conflicting goals and interests, which raises the need for fairness and popularity bias considerations. GNN-based recommendation methods also face the challenges of unfairness and popularity bias, and their normalization and aggregation processes suffer from these challenges. In this paper, we propose a fair GNN-based recommender system, called HetroFair, to improve item-side fairness. HetroFair uses two separate components to generate fairness-aware embeddings: i) Fairness-aware attention, which incorporates the dot product in the normalization process of GNNs to decrease the effect of nodes' degrees. ii) Heterophily feature weighting, to assign distinct weights to different features during the aggregation process. To evaluate the effectiveness of HetroFair, we conduct extensive experiments over six real-world datasets. Our experimental results reveal that HetroFair not only alleviates unfairness and popularity bias on the item side but also achieves superior accuracy on the user side. Our implementation is publicly available at https://github.com/NematGH/HetroFair.
Authors: Ling-Qi Zhang, Zahra Kadkhodaie, Eero P. Simoncelli, David H. Brainard
Abstract: We examine the problem of selecting a small set of linear measurements for reconstructing high-dimensional signals. Well-established methods for optimizing such measurements include principal component analysis (PCA), independent component analysis (ICA) and compressed sensing (CS) based on random projections, all of which rely on axis- or subspace-aligned statistical characterization of the signal source. However, many naturally occurring signals, including photographic images, contain richer statistical structure. To exploit such structure, we introduce a general method for obtaining an optimized set of linear measurements for efficient image reconstruction, where the signal statistics are expressed by the prior implicit in a neural network trained to perform denoising (known as a "diffusion model"). We demonstrate that the optimal measurements derived for two natural image datasets differ from those of PCA, ICA, or CS, and result in substantially lower mean squared reconstruction error. Interestingly, the marginal distributions of the measurement values are asymmetrical (skewed), substantially more so than those of previous methods. We also find that optimizing with respect to perceptual loss, as quantified by structural similarity (SSIM), leads to measurements different from those obtained when optimizing for MSE. Our results highlight the importance of incorporating the specific statistical regularities of natural signals when designing effective linear measurements.
Authors: Chang Wu, Yuang Chen, Hancheng Lu
Abstract: More refined resource management and Quality of Service (QoS) provisioning is a critical goal of wireless communication technologies. In this paper, we propose a novel Business-Centric Network (BCN) aimed at enabling scalable QoS provisioning, based on a cross-layer framework that captures the relationship between application, transport parameters, and channels. We investigate both continuous flow and event-driven flow models, presenting key QoS metrics such as throughput, delay, and reliability. By jointly considering power and bandwidth allocation, transmission parameters, and AP network topology across layers, we optimize weighted resource efficiency with statistical QoS provisioning. To address the coupling among parameters, we propose a novel deep reinforcement learning (DRL) framework, which is Collaborative Optimization among Heterogeneous Actors with Experience Sharing (COHA-ES). Power and sub-channel (SC) Actors representing multiple APs are jointly optimized under the unified guidance of a common critic. Additionally, we introduce a novel multithreaded experience-sharing mechanism to accelerate training and enhance rewards. Extensive comparative experiments validate the effectiveness of our DRL framework in terms of convergence and efficiency. Moreover, comparative analyses demonstrate the comprehensive advantages of the BCN structure in enhancing both spectral and energy efficiency.
Authors: Jiawei Lian, Shaohui Mei, Xiaofei Wang, Yi Wang, Lefan Wang, Yingjie Lu, Mingyang Ma, Lap-Pui Chau
Abstract: It has been widely substantiated that deep neural networks (DNNs) are susceptible and vulnerable to adversarial perturbations. Existing studies mainly focus on performing attacks by corrupting targeted objects (physical attack) or images (digital attack), which is intuitively acceptable and understandable in terms of the attack's effectiveness. In contrast, our focus lies in conducting background adversarial attacks in both digital and physical domains, without causing any disruptions to the targeted objects themselves. Specifically, an effective background adversarial attack framework is proposed to attack anything, by which the attack efficacy generalizes well between diverse objects, models, and tasks. Technically, we approach the background adversarial attack as an iterative optimization problem, analogous to the process of DNN learning. Besides, we offer a theoretical demonstration of its convergence under a set of mild but sufficient conditions. To strengthen the attack efficacy and transferability, we propose a new ensemble strategy tailored for adversarial perturbations and introduce an improved smooth constraint for the seamless connection of integrated perturbations. We conduct comprehensive and rigorous experiments in both digital and physical domains across various objects, models, and tasks, demonstrating the effectiveness of attacking anything of the proposed method. The findings of this research substantiate the significant discrepancy between human and machine vision on the value of background variations, which play a far more critical role than previously recognized, necessitating a reevaluation of the robustness and reliability of DNNs. The code will be publicly available at https://github.com/JiaweiLian/Attack_Anything
Authors: Kota Dohi, Aoi Ito, Harsh Purohit, Tomoya Nishida, Takashi Endo, Yohei Kawaguchi
Abstract: Due to scarcity of time-series data annotated with descriptive texts, training a model to generate descriptive texts for time-series data is challenging. In this study, we propose a method to systematically generate domain-independent descriptive texts from time-series data. We identify two distinct approaches for creating pairs of time-series data and descriptive texts: the forward approach and the backward approach. By implementing the novel backward approach, we create the Temporal Automated Captions for Observations (TACO) dataset. Experimental results demonstrate that a contrastive learning based model trained using the TACO dataset is capable of generating descriptive texts for time-series data in novel domains.
Authors: Pura Peetathawatchai, Wei-Ning Chen, Berivan Isik, Sanmi Koyejo, Albert No
Abstract: Personalizing large-scale diffusion models poses serious privacy risks, especially when adapting to small, sensitive datasets. A common approach is to fine-tune the model using differentially private stochastic gradient descent (DP-SGD), but this suffers from severe utility degradation due to the high noise needed for privacy, particularly in the small data regime. We propose an alternative that leverages Textual Inversion (TI), which learns an embedding vector for an image or set of images, to enable adaptation under differential privacy (DP) constraints. Our approach, Differentially Private Aggregation via Textual Inversion (DPAgg-TI), adds calibrated noise to the aggregation of per-image embeddings to ensure formal DP guarantees while preserving high output fidelity. We show that DPAgg-TI outperforms DP-SGD finetuning in both utility and robustness under the same privacy budget, achieving results closely matching the non-private baseline on style adaptation tasks using private artwork from a single artist and Paris 2024 Olympic pictograms. In contrast, DP-SGD fails to generate meaningful outputs in this setting.
Authors: Xin Yang, Omid Ardakanian
Abstract: Sensor data collected by Internet of Things (IoT) devices can reveal sensitive personal information about individuals, raising significant privacy concerns when shared with semi-trusted service providers, as they may extract this information using machine learning models. Data obfuscation empowered by generative models is a promising approach to generate synthetic data such that useful information contained in the original data is preserved while sensitive information is obscured. This newly generated data will then be shared with service providers instead of the original sensor data. In this work, we propose PrivDiffuser, a novel data obfuscation technique based on a denoising diffusion model that achieves a superior trade-off between data utility and privacy by incorporating effective guidance techniques. Specifically, we extract latent representations that contain information about public and private attributes from sensor data to guide the diffusion model, and impose mutual information-based regularization when learning the latent representations to alleviate the entanglement of public and private attributes, thereby increasing the effectiveness of guidance. Evaluation on three real-world datasets containing different sensing modalities reveals that PrivDiffuser yields a better privacy-utility trade-off than the state-of-the-art in data obfuscation, decreasing the utility loss by up to $1.81\%$ and the privacy loss by up to $3.42\%$. Moreover, compared with existing obfuscation approaches, PrivDiffuser offers the unique benefit of allowing users with diverse privacy needs to protect their privacy without having to retrain the generative model.
Authors: Marta Gentiloni-Silveri, Antonio Ocello
Abstract: Score-based Generative Models (SGMs) aim to sample from a target distribution by learning score functions using samples perturbed by Gaussian noise. Existing convergence bounds for SGMs in the $\mathcal{W}_2$-distance rely on stringent assumptions about the data distribution. In this work, we present a novel framework for analyzing $\mathcal{W}_2$-convergence in SGMs, significantly relaxing traditional assumptions such as log-concavity and score regularity. Leveraging the regularization properties of the Ornstein--Uhlenbeck (OU) process, we show that weak log-concavity of the data distribution evolves into log-concavity over time. This transition is rigorously quantified through a PDE-based analysis of the Hamilton--Jacobi--Bellman equation governing the log-density of the forward process. Moreover, we establish that the drift of the time-reversed OU process alternates between contractive and non-contractive regimes, reflecting the dynamics of concavity. Our approach circumvents the need for stringent regularity conditions on the score function and its estimators, relying instead on milder, more practical assumptions. We demonstrate the wide applicability of this framework through explicit computations on Gaussian mixture models, illustrating its versatility and potential for broader classes of data distributions.
Authors: Krit Tangsongcharoen, Teerachote Pakornchote, Chayanon Atthapak, Natthaphon Choomphon-anomakhun, Annop Ektarawong, Bj\"orn Alling, Christopher Sutton, Thiti Bovornratanaraks, Thiparat Chotibut
Abstract: Determining whether a candidate crystalline material is thermodynamically stable depends on identifying its true ground-state structure, a central challenge in computational materials science. We introduce CrystalGRW, a diffusion-based generative model on Riemannian manifolds that proposes novel crystal configurations and can predict stable phases validated by density functional theory. The crystal properties, such as fractional coordinates, atomic types, and lattice matrices, are represented on suitable Riemannian manifolds, ensuring that new predictions generated through the diffusion process preserve the periodicity of crystal structures. We incorporate an equivariant graph neural network to also account for rotational and translational symmetries during the generation process. CrystalGRW demonstrates the ability to generate realistic crystal structures that are close to their ground states with accuracy comparable to existing models, while also enabling conditional control, such as specifying a desired crystallographic point group. These features help accelerate materials discovery and inverse design by offering stable, symmetry-consistent crystal candidates for experimental validation.
Authors: Guangjin Pan, Yuan Gao, Yilin Gao, Wenjun Yu, Zhiyong Zhong, Xiaoyu Yang, Xinyu Guo, Shugong Xu
Abstract: Wireless positioning technologies hold significant value for applications in autonomous driving, extended reality (XR), unmanned aerial vehicles (UAVs), and more. With the advancement of artificial intelligence (AI), leveraging AI to enhance positioning accuracy and robustness has emerged as a field full of potential. Driven by the requirements and functionalities defined in the 3rd Generation Partnership Project (3GPP) standards, AI/machine learning (ML)-based cellular positioning is becoming a key technology to overcome the limitations of traditional methods. This paper presents a comprehensive survey of AI-driven cellular positioning. We begin by reviewing the fundamentals of wireless positioning and AI models, analyzing their respective challenges and synergies. We provide a comprehensive review of the evolution of 3GPP positioning standards, with a focus on the integration of AI/ML in current and upcoming standard releases. Guided by the 3GPP-defined taxonomy, we categorize and summarize state-of-the-art (SOTA) research into two major classes: AI/ML-assisted positioning and direct AI/ML-based positioning. The former includes line-of-sight (LOS)/non-line-of-sight (NLOS) detection, time of arrival (TOA)/time difference of arrival (TDOA) estimation, and angle prediction; the latter encompasses fingerprinting, knowledge-assisted learning, and channel charting. Furthermore, we review representative public datasets and conduct performance evaluations of AI-based positioning algorithms using these datasets. Finally, we conclude by summarizing the challenges and opportunities of AI-driven wireless positioning.
Authors: Haoran Sun, Yusen Wu, Peng Wang, Wei Chen, Yukun Cheng, Xiaotie Deng, Xu Chu
Abstract: Game theory is a foundational framework for analyzing strategic interactions, and its intersection with large language models (LLMs) is a rapidly growing field. However, existing surveys mainly focus narrowly on using game theory to evaluate LLM behavior. This paper provides the first comprehensive survey of the bidirectional relationship between Game Theory and LLMs. We propose a novel taxonomy that categorizes the research in this intersection into four distinct perspectives: (1) evaluating LLMs in game-based scenarios; (2) improving LLMs using game-theoretic concepts for better interpretability and alignment; (3) modeling the competitive landscape of LLM development and its societal impact; and (4) leveraging LLMs to advance game models and to solve corresponding game theory problems. Furthermore, we identify key challenges and outline future research directions. By systematically investigating this interdisciplinary landscape, our survey highlights the mutual influence of game theory and LLMs, fostering progress at the intersection of these fields.
Authors: Zhen Zou, Feng Zhao
Abstract: Diffusion Transformer (DiT) has exhibited impressive generation capabilities but faces great challenges due to its high computational complexity. To address this issue, various methods, notably feature caching, have been introduced. However, these approaches focus on aligning non-cache diffusion without analyzing why caching damage the generation processes. In this paper, we first confirm that the cache greatly amplifies the exposure bias, resulting in a decline in the generation quality. However, directly applying noise scaling is challenging for this issue due to the non-smoothness of exposure bias. We found that this phenomenon stems from the mismatch between its frequency response characteristics and the simple cache of Attention and MLP. Since these two components exhibit unique preferences for frequency signals, which provides us with a caching strategy to separate Attention and MLP to achieve an enhanced fit of exposure bias and reduce it. Based on this, we introduced FEB-Cache, a joint caching strategy that aligns with the non-exposed bias diffusion process (which gives us a higher performance cap) of caching Attention and MLP based on the frequency-guided cache table. Our approach combines a comprehensive understanding of the caching mechanism and offers a new perspective on leveraging caching to accelerate the diffusion process. Empirical results indicate that FEB-Cache optimizes model performance while concurrently facilitating acceleration.
Authors: Xiaoyu Huang, Takara Truong, Yunbo Zhang, Fangzhou Yu, Jean Pierre Sleiman, Jessica Hodgins, Koushil Sreenath, Farbod Farshidian
Abstract: We present Diffuse-CLoC, a guided diffusion framework for physics-based look-ahead control that enables intuitive, steerable, and physically realistic motion generation. While existing kinematics motion generation with diffusion models offer intuitive steering capabilities with inference-time conditioning, they often fail to produce physically viable motions. In contrast, recent diffusion-based control policies have shown promise in generating physically realizable motion sequences, but the lack of kinematics prediction limits their steerability. Diffuse-CLoC addresses these challenges through a key insight: modeling the joint distribution of states and actions within a single diffusion model makes action generation steerable by conditioning it on the predicted states. This approach allows us to leverage established conditioning techniques from kinematic motion generation while producing physically realistic motions. As a result, we achieve planning capabilities without the need for a high-level planner. Our method handles a diverse set of unseen long-horizon downstream tasks through a single pre-trained model, including static and dynamic obstacle avoidance, motion in-betweening, and task-space control. Experimental results show that our method significantly outperforms the traditional hierarchical framework of high-level motion diffusion and low-level tracking.
Authors: Mari Ashiga, Wei Jie, Fan Wu, Vardan Voskanyan, Fateme Dinmohammadi, Paul Brookes, Jingzhi Gong, Zheng Wang
Abstract: Generative Pretrained Transformers (GPTs) are foundational Large Language Models (LLMs) for text generation. However, individual LLMs often produce inconsistent outputs and exhibit biases, limiting their representation of diverse language patterns. The closed-source nature of many powerful LLMs further restricts industry applications due to data privacy concerns. Inspired by successes in text generation, LLM ensemble techniques are now increasingly explored for code generation. This article reviews these emerging ensemble approaches to enhance understanding, encourage further research, and promote practical implementation in both text and code generation. We categorize LLM ensembles into seven main methods - weight merging, knowledge fusion, mixture-of-experts, reward ensemble, output ensemble, routing, and cascading - analyzing capabilities of those approaches. Our findings highlight key benefits such as improved diversity representation, enhanced output quality, and greater application flexibility. These insights aid model selection for real-world tasks and crucially, lay groundwork for extending ensemble strategies to multimodal LLMs.
Authors: Arslan Mazitov, Filippo Bigi, Matthias Kellner, Paolo Pegolo, Davide Tisi, Guillaume Fraux, Sergey Pozdnyakov, Philip Loche, Michele Ceriotti
Abstract: Machine-learning interatomic potentials (MLIPs) have greatly extended the reach of atomic-scale simulations, offering the accuracy of first-principles calculations at a fraction of the cost. Leveraging large quantum mechanical databases and expressive architectures, recent ''universal'' models deliver qualitative accuracy across the periodic table but are often biased toward low-energy configurations. We introduce PET-MAD, a generally applicable MLIP trained on a dataset combining stable inorganic and organic solids, systematically modified to enhance atomic diversity. Using a moderate but highly-consistent level of electronic-structure theory, we assess PET-MAD's accuracy on established benchmarks and advanced simulations of six materials. Despite the small training set and lightweight architecture, PET-MAD is competitive with state-of-the-art MLIPs for inorganic solids, while also being reliable for molecules, organic materials, and surfaces. It is stable and fast, enabling the near-quantitative study of thermal and quantum mechanical fluctuations, functional properties, and phase transitions out of the box. It can be efficiently fine-tuned to deliver full quantum mechanical accuracy with a minimal number of targeted calculations.
Authors: Afrah Farea, Saiful Khan, Mustafa Serdar Celebi
Abstract: Physics-informed neural networks (PINNs) have emerged as promising methods for solving partial differential equations (PDEs) by embedding physical laws within neural architectures. However, these classical approaches often require a large number of parameters to achieve reasonable accuracy, particularly for complex PDEs. In this paper, we present a quantum-classical physics-informed neural network (QCPINN) that combines quantum and classical components, allowing us to solve PDEs with significantly fewer parameters while maintaining comparable accuracy and convergence to classical PINNs. We systematically evaluated two quantum circuit architectures across various configurations on five benchmark PDEs to identify optimal QCPINN designs. Our results demonstrate that the QCPINN achieves stable convergence and comparable accuracy, while requiring approximately 10\% of the trainable parameters used in classical approaches. It also results in a 40\% reduction in the relative error $L_2$ for the convection-diffusion equation. These findings demonstrate the potential of parameter efficiency as a measurable quantum advantage in physics-informed machine learning, significantly reducing model complexity while preserving solution quality. This approach presents a promising solution to the computational challenges associated with solving PDEs.
Authors: Chenxi Wang, Jizhan Fang, Xiang Chen, Bozhong Tian, Ziwen Xu, Huajun Chen, Ningyu Zhang
Abstract: Recent advancements in Large Multimodal Models (LMMs) have shown promise in Autonomous Driving Systems (ADS). However, their direct application to ADS is hindered by challenges such as misunderstanding of traffic knowledge, complex road conditions, and diverse states of vehicle. To address these challenges, we propose the use of Knowledge Editing, which enables targeted modifications to a model's behavior without the need for full retraining. Meanwhile, we introduce ADS-Edit, a multimodal knowledge editing dataset specifically designed for ADS, which includes various real-world scenarios, multiple data types, and comprehensive evaluation metrics. We conduct comprehensive experiments and derive several interesting conclusions. We hope that our work will contribute to the further advancement of knowledge editing applications in the field of autonomous driving. Code and data are available in https://github.com/zjunlp/EasyEdit/blob/main/examples/ADSEdit.md.
URLs: https://github.com/zjunlp/EasyEdit/blob/main/examples/ADSEdit.md.
Authors: Jannik Endres, Oliver Hahn, Charles Corbi\`ere, Simone Schaub-Meyer, Stefan Roth, Alexandre Alahi
Abstract: Omnidirectional depth perception is essential for mobile robotics applications that require scene understanding across a full 360{\deg} field of view. Camera-based setups offer a cost-effective option by using stereo depth estimation to generate dense, high-resolution depth maps without relying on expensive active sensing. However, existing omnidirectional stereo matching approaches achieve only limited depth accuracy across diverse environments, depth ranges, and lighting conditions, due to the scarcity of real-world data. We present DFI-OmniStereo, a novel omnidirectional stereo matching method that leverages a large-scale pre-trained foundation model for relative monocular depth estimation within an iterative optimization-based stereo matching architecture. We introduce a dedicated two-stage training strategy to utilize the relative monocular depth features for our omnidirectional stereo matching before scale-invariant fine-tuning. DFI-OmniStereo achieves state-of-the-art results on the real-world Helvipad dataset, reducing disparity MAE by approximately 16% compared to the previous best omnidirectional stereo method.
Authors: Kyungmin Choi, JaKeoung Koo, Stephen McLaughlin, Abderrahim Halimi
Abstract: Single-photon Lidar imaging offers a significant advantage in 3D imaging due to its high resolution and long-range capabilities, however it is challenging to apply in noisy environments with multiple targets per pixel. To tackle these challenges, several methods have been proposed. Statistical methods demonstrate interpretability on the inferred parameters, but they are often limited in their ability to handle complex scenes. Deep learning-based methods have shown superior performance in terms of accuracy and robustness, but they lack interpretability or they are limited to a single-peak per pixel. In this paper, we propose a deep unrolling algorithm for dual-peak single-photon Lidar imaging. We introduce a hierarchical Bayesian model for multiple targets and propose a neural network that unrolls the underlying statistical method. To support multiple targets, we adopt a dual depth maps representation and exploit geometric deep learning to extract features from the point cloud. The proposed method takes advantages of statistical methods and learning-based methods in terms of accuracy and quantifying uncertainty. The experimental results on synthetic and real data demonstrate the competitive performance when compared to existing methods, while also providing uncertainty information.
Authors: Dongyang Fan, Vinko Sabol\v{c}ec, Matin Ansaripour, Ayush Kumar Tarun, Martin Jaggi, Antoine Bosselut, Imanol Schlag
Abstract: The increasing adoption of web crawling opt-outs by copyright holders of online content raises critical questions about the impact of data compliance on large language model (LLM) performance. However, little is known about how these restrictions (and the resultant filtering of pretraining datasets) affect the capabilities of models trained using these corpora. In this work, we conceptualize this effect as the $\textit{data compliance gap}$ (DCG), which quantifies the performance difference between models trained on datasets that comply with web crawling opt-outs, and those that do not. We measure the data compliance gap in two settings: pretraining models from scratch and continual pretraining from existing compliant models (simulating a setting where copyrighted data could be integrated later in pretraining). Our experiments with 1.5B models show that, as of January 2025, compliance with web data opt-outs does not degrade general knowledge acquisition (close to 0\% DCG). However, in specialized domains such as biomedical research, excluding major publishers leads to performance declines. These findings suggest that while general-purpose LLMs can be trained to perform equally well using fully open data, performance in specialized domains may benefit from access to high-quality copyrighted sources later in training. Our study provides empirical insights into the long-debated trade-off between data compliance and downstream model performance, informing future discussions on AI training practices and policy decisions. Our website is available at https://data-compliance.github.io/.
Authors: Shahriar Noroozizadeh, Jeremy C. Weiss
Abstract: Clinical case reports and discharge summaries may be the most complete and accurate summarization of patient encounters, yet they are finalized, i.e., timestamped after the encounter. Complementary data structured streams become available sooner but suffer from incompleteness. To train models and algorithms on more complete and temporally fine-grained data, we construct a pipeline to phenotype, extract, and annotate time-localized findings within case reports using large language models. We apply our pipeline to generate an open-access textual time series corpus for Sepsis-3 comprising 2,139 case reports from the Pubmed-Open Access (PMOA) Subset. To validate our system, we apply it on PMOA and timeline annotations from I2B2/MIMIC-IV and compare the results to physician-expert annotations. We show high recovery rates of clinical findings (event match rates: O1-preview--0.755, Llama 3.3 70B Instruct--0.753) and strong temporal ordering (concordance: O1-preview--0.932, Llama 3.3 70B Instruct--0.932). Our work characterizes the ability of LLMs to time-localize clinical findings in text, illustrating the limitations of LLM use for temporal reconstruction and providing several potential avenues of improvement via multimodal integration.
Authors: Anamika Lochab, Ruqi Zhang
Abstract: Reward models (RMs) are essential for aligning Large Language Models (LLMs) with human preferences. However, they often struggle with capturing complex human preferences and generalizing to unseen data. To address these challenges, we introduce Energy-Based Reward Model (EBRM), a lightweight post-hoc refinement framework that enhances RM robustness and generalization. EBRM models the reward distribution explicitly, capturing uncertainty in human preferences and mitigating the impact of noisy or misaligned annotations. It achieves this through conflict-aware data filtering, label-noise-aware contrastive training, and hybrid initialization. Notably, EBRM enhances RMs without retraining, making it computationally efficient and adaptable across different models and tasks. Empirical evaluations on RM benchmarks demonstrate significant improvements in both robustness and generalization, achieving up to a 5.97% improvement in safety-critical alignment tasks compared to standard RMs. Furthermore, reinforcement learning experiments confirm that our refined rewards enhance alignment quality, effectively delaying reward hacking. These results demonstrate our approach as a scalable and effective enhancement for existing RMs and alignment pipelines. The code is available at EBRM.
Authors: Tingfeng Lan, Yusen Wu, Bin Ma, Zhaoyuan Su, Rui Yang, Tekin Bicer, Masahiro Tanaka, Olatunji Ruwase, Dong Li, Yue Cheng
Abstract: Fine-tuning large language models (LLMs) often exceeds GPU memory limits, prompting systems to offload model states to CPU memory. However, existing offloaded training frameworks like ZeRO-Offload treat all parameters equally and update the full model on the CPU, causing severe GPU stalls, where fast, expensive GPUs sit idle waiting for slow CPU updates and limited-bandwidth PCIe transfers. We present ZenFlow, a new offloading framework that prioritizes important parameters and decouples updates between GPU and CPU. ZenFlow performs in-place updates of important gradients on GPU, while asynchronously offloading and accumulating less important ones on CPU, fully overlapping CPU work with GPU computation. To scale across GPUs, ZenFlow introduces a lightweight gradient selection method that exploits a novel spatial and temporal locality property of important gradients, avoiding costly global synchronization. ZenFlow achieves up to 5x end-to-end speedup, 2x lower PCIe traffic, and reduces GPU stalls by over 85 percent, all while preserving accuracy.
Authors: Shibo Hong, Jiahao Ying, Haiyuan Liang, Mengdi Zhang, Jun Kuang, Jiazheng Zhang, Yixin Cao
Abstract: Evaluating the open-ended outputs of Large Multimodal Models has become a bottleneck as model capabilities, task diversity, and modality rapidly expand. Existing "LLM-as-a-Judge" evaluators are typically narrow in specific tasks and aspects. In this paper, we argue that, on one hand, based on the interconnected nature of aspects, learning specific aspects can generalize to unseen aspects; on the other hand, jointly learning to assess multiple visual aspects and tasks may foster a synergistic effect. To this end, we propose UFEval, the first unified fine-grained evaluator with task and aspect generalization for four evaluation tasks -- Natural Language Generation, Image Understanding, Image Generation, and Interleaved Text-and-Image Generation. Specifically, (1) We first construct a hierarchical aspect taxonomy encompassing 112 distinct aspects across the aforementioned four tasks. (2) Then, building upon this taxonomy, we create FRABench, a fine-grained evaluation dataset comprising 60.4k pairwise samples with 325k evaluation labels obtained from a combination of human and GPT-4o annotations. FRABench provides a large-scale, multi-modal, and aspect-level resource for training and testing evaluators. (3) Finally, leveraging FRABench, we develop UFEval, a unified fine-grained evaluator. Experiments show that learning on specific aspects enables UFEval to generalize to unseen aspects, and joint learning to assess diverse tasks and aspects can lead to substantial mutual benefits.
Authors: Kunal Sawarkar, Shivam R. Solanki, Abhilasha Mangal
Abstract: Retrieval-Augmented Generation (RAG) struggles with domain-specific enterprise datasets, often isolated behind firewalls and rich in complex, specialized terminology unseen by LLMs during pre-training. Semantic variability across domains like medicine, networking, or law hampers RAG's context precision, while fine-tuning solutions are costly, slow, and lack generalization as new data emerges. Achieving zero-shot precision with retrievers without fine-tuning still remains a key challenge. We introduce 'MetaGen Blended RAG', a novel enterprise search approach that enhances semantic retrievers through a metadata generation pipeline and hybrid query indexes using dense and sparse vectors. By leveraging key concepts, topics, and acronyms, our method creates metadata-enriched semantic indexes and boosted hybrid queries, delivering robust, scalable performance without fine-tuning. On the biomedical PubMedQA dataset, MetaGen Blended RAG achieves 82% retrieval accuracy and 77% RAG accuracy, surpassing all prior zero-shot RAG benchmarks and even rivaling fine-tuned models on that dataset, while also excelling on datasets like SQuAD and NQ. This approach redefines enterprise search using a new approach to building semantic retrievers with unmatched generalization across specialized domains.
Authors: Mario Leiva, Noel Ngu, Joshua Shay Kricheli, Aditya Taparia, Ransalu Senanayake, Paulo Shakarian, Nathaniel Bastian, John Corcoran, Gerardo Simari
Abstract: The deployment of pre-trained perception models in novel environments often leads to performance degradation due to distributional shifts. Although recent artificial intelligence approaches for metacognition use logical rules to characterize and filter model errors, improving precision often comes at the cost of reduced recall. This paper addresses the hypothesis that leveraging multiple pre-trained models can mitigate this recall reduction. We formulate the challenge of identifying and managing conflicting predictions from various models as a consistency-based abduction problem, building on the idea of abductive learning (ABL) but applying it to test-time instead of training. The input predictions and the learned error detection rules derived from each model are encoded in a logic program. We then seek an abductive explanation--a subset of model predictions--that maximizes prediction coverage while ensuring the rate of logical inconsistencies (derived from domain constraints) remains below a specified threshold. We propose two algorithms for this knowledge representation task: an exact method based on Integer Programming (IP) and an efficient Heuristic Search (HS). Through extensive experiments on a simulated aerial imagery dataset featuring controlled, complex distributional shifts, we demonstrate that our abduction-based framework outperforms individual models and standard ensemble baselines, achieving, for instance, average relative improvements of approximately 13.6\% in F1-score and 16.6\% in accuracy across 15 diverse test datasets when compared to the best individual model. Our results validate the use of consistency-based abduction as an effective mechanism to robustly integrate knowledge from multiple imperfect models in challenging, novel scenarios.
Authors: Deepak Pandita, Tharindu Cyril Weerasooriya, Ankit Parag Shah, Isabelle Diana May-Xin Ng, Christopher M. Homan, Wei Wei
Abstract: Agentic workflows, where multiple AI agents collaborate to accomplish complex tasks like reasoning or planning, play a substantial role in many cutting-edge commercial applications, and continue to fascinate researchers across nearly all fields for their potential to accomplish expensive, complex tasks that, until recently, only humans have been trusted to do. These workflows depend critically on the prompts used to provide the roles models play in such workflows. Poorly designed prompts that fail even slightly to guide individual agents can lead to sub-optimal performance that may snowball within a system of agents, limiting their reliability and scalability. To address this important problem of inference-time prompt optimization, we introduce ProRefine, an innovative inference-time optimization method that uses an agentic loop of LLMs to generate and apply textual feedback. ProRefine dynamically refines prompts for multi-step reasoning tasks without additional training or ground truth labels. Evaluated on five benchmark mathematical reasoning datasets, ProRefine significantly surpasses zero-shot Chain-of-Thought baselines by 3 to 37 percentage points. This approach not only boosts accuracy but also allows smaller models to approach the performance of their larger counterparts. This highlights its potential for building more cost-effective and powerful hybrid AI systems, thereby democratizing access to high-performing AI.
Authors: Kangwei Liu, Siyuan Cheng, Bozhong Tian, Xiaozhuan Liang, Yuyang Yin, Meng Han, Ningyu Zhang, Bryan Hooi, Xi Chen, Shumin Deng
Abstract: Large language models (LLMs) have been increasingly applied to automated harmful content detection tasks, assisting moderators in identifying policy violations and improving the overall efficiency and accuracy of content review. However, existing resources for harmful content detection are predominantly focused on English, with Chinese datasets remaining scarce and often limited in scope. We present a comprehensive, professionally annotated benchmark for Chinese content harm detection, which covers six representative categories and is constructed entirely from real-world data. Our annotation process further yields a knowledge rule base that provides explicit expert knowledge to assist LLMs in Chinese harmful content detection. In addition, we propose a knowledge-augmented baseline that integrates both human-annotated knowledge rules and implicit knowledge from large language models, enabling smaller models to achieve performance comparable to state-of-the-art LLMs. Code and data are available at https://github.com/zjunlp/ChineseHarm-bench.
Authors: Yuqi Zhu, Yi Zhong, Jintian Zhang, Ziheng Zhang, Shuofei Qiao, Yujie Luo, Lun Du, Da Zheng, Ningyu Zhang, Huajun Chen
Abstract: Large Language Models (LLMs) hold promise in automating data analysis tasks, yet open-source models face significant limitations in these kinds of reasoning-intensive scenarios. In this work, we investigate strategies to enhance the data analysis capabilities of open-source LLMs. By curating a seed dataset of diverse, realistic scenarios, we evaluate model behavior across three core dimensions: data understanding, code generation, and strategic planning. Our analysis reveals three key findings: (1) Strategic planning quality serves as the primary determinant of model performance; (2) Interaction design and task complexity significantly influence reasoning capabilities; (3) Data quality demonstrates a greater impact than diversity in achieving optimal performance. We leverage these insights to develop a data synthesis methodology, demonstrating significant improvements in open-source LLMs' analytical reasoning capabilities. Code is available at https://github.com/zjunlp/DataMind.
Authors: Yuchi Tang, I\~naki Esnaola, George Panoutsos
Abstract: Existing post-hoc model-agnostic methods generate external explanations for opaque models, primarily by locally attributing the model output to its input features. However, they often lack an explicit and systematic framework for quantifying the contribution of individual features. Building on the Taylor expansion framework introduced by Deng et al. (2024) to unify existing local attribution methods, we propose a rigorous set of postulates -- "precision", "federation", and "zero-discrepancy" -- to govern Taylor term-specific attribution. Guided by these postulates, we introduce TaylorPODA (Taylor expansion-derived imPortance-Order aDapted Attribution), which incorporates an additional "adaptation" property. This property enables alignment with task-specific goals, especially in post-hoc settings lacking ground-truth explanations. Empirical evaluations demonstrate that TaylorPODA achieves competitive results against baseline methods, providing principled and visualization-friendly explanations. This work enhances the trustworthy deployment of opaque models by offering explanations with stronger theoretical grounding.
Authors: Goeric Huybrechts, Srikanth Ronanki, Sai Muralidhar Jayanthi, Jack Fitzgerald, Srinivasan Veeravanallur
Abstract: The proliferation of multimodal Large Language Models has significantly advanced the ability to analyze and understand complex data inputs from different modalities. However, the processing of long documents remains under-explored, largely due to a lack of suitable benchmarks. To address this, we introduce Document Haystack, a comprehensive benchmark designed to evaluate the performance of Vision Language Models (VLMs) on long, visually complex documents. Document Haystack features documents ranging from 5 to 200 pages and strategically inserts pure text or multimodal text+image "needles" at various depths within the documents to challenge VLMs' retrieval capabilities. Comprising 400 document variants and a total of 8,250 questions, it is supported by an objective, automated evaluation framework. We detail the construction and characteristics of the Document Haystack dataset, present results from prominent VLMs and discuss potential research avenues in this area.
Authors: Lechen Feng, Xun Li, Yuan-Hua Ni
Abstract: This paper develops a unified nonconvex optimization framework for the design of group-sparse feedback controllers in infinite-horizon linear-quadratic (LQ) problems. We address two prominent extensions of the classical LQ problem: the distributed LQ problem with fixed communication topology (DFT-LQ) and the sparse feedback LQ problem (SF-LQ), both of which are motivated by the need for scalable and structure-aware control in large-scale systems. Unlike existing approaches that rely on convex relaxations or are limited to block-diagonal structures, we directly formulate the controller synthesis as a finite-dimensional nonconvex optimization problem with group $\ell_0$-norm regularization, capturing general sparsity patterns. We establish a connection between DFT-LQ and SF-LQ problems, showing that both can be addressed within our unified framework. Furthermore, we propose a penalty-based proximal alternating linearized minimization (PALM) algorithm and provide a rigorous convergence analysis under mild assumptions, overcoming the lack of coercivity in the objective function. The proposed method admits efficient solvers for all subproblems and guarantees global convergence to critical points. Our results fill a key gap in the literature by enabling the direct design of group-sparse feedback gains with theoretical guarantees, without resorting to convex surrogates or restrictive structural assumptions.
Authors: Minje Park, Jeonghwa Lim, Taehyung Yu, Sunghoon Joo
Abstract: Electrocardiogram (ECG) delineation, the segmentation of meaningful waveform features, is critical for clinical diagnosis. Despite recent advances using deep learning, progress has been limited by the scarcity of publicly available annotated datasets. Semi-supervised learning presents a promising solution by leveraging abundant unlabeled ECG data. In this study, we present SemiSegECG, the first systematic benchmark for semi-supervised semantic segmentation (SemiSeg) in ECG delineation. We curated and unified multiple public datasets, including previously underused sources, to support robust and diverse evaluation. We adopted five representative SemiSeg algorithms from computer vision, implemented them on two different architectures: the convolutional network and the transformer, and evaluated them in two different settings: in-domain and cross-domain. Additionally, we propose ECG-specific training configurations and augmentation strategies and introduce a standardized evaluation framework. Our results show that the transformer outperforms the convolutional network in semi-supervised ECG delineation. We anticipate that SemiSegECG will serve as a foundation for advancing semi-supervised ECG delineation methods and will facilitate further research in this domain.
Authors: Lechen Feng, Xun Li, Yuan-Hua Ni
Abstract: This work is a companion paper of [8], where the distributed linear-quadratic problem with fixed communication topology (DFT-LQ) and the sparse feedback LQ problem (SF-LQ) are formulated into a nonsmooth and nonconvex optimization problem with affine constraints. Moreover, a penalty approach is considered in \cite{feng-part1}, and the PALM (proximal alternating linearized minimization) algorithm is studied with convergence and complexity analysis. In this paper, we aim to address the inherent drawbacks of the penalty approach, such as the challenge of tuning the penalty parameter and the risk of introducing spurious stationary points. Specifically, we first reformulate the SF-LQ problem and the DFT-LQ problem from an epi-composition function perspective, aiming to solve the constrained problem directly. Then, from a theoretical viewpoint, we revisit the alternating direction method of multipliers (ADMM) and establish its convergence to the set of cluster points under certain assumptions. When these assumptions do not hold, we can effectively utilize alternative approaches combining subgradient descent with Difference-of-Convex relaxation methods. In summary, our results enable the direct design of group-sparse feedback gains with theoretical guarantees, without resorting to convex surrogates, restrictive structural assumptions, or penalty formulations that incorporate constraints into the cost function.
Authors: Santosh Patapati, Trisanth Srinivasan, Murari Ambati
Abstract: Autonomous cars need geometric accuracy and semantic understanding to navigate complex environments, yet most stacks handle them separately. We present XYZ-Drive, a single vision-language model that reads a front-camera frame, a 25m $\times$ 25m overhead map, and the next waypoint, then outputs steering and speed. A lightweight goal-centered cross-attention layer lets waypoint tokens highlight relevant image and map patches, supporting both action and textual explanations, before the fused tokens enter a partially fine-tuned LLaMA-3.2 11B model. On the MD-NEX Outdoor-Driving benchmark XYZ-Drive attains 95% success and 0.80 Success weighted by Path Length (SPL), surpassing PhysNav-DG by 15%. and halving collisions, all while significantly improving efficiency by using only a single branch. Sixteen ablations explain the gains. Removing any modality (vision, waypoint, map) drops success by up to 11%, confirming their complementary roles and rich connections. Replacing goal-centered attention with simple concatenation cuts 3% in performance, showing query-based fusion injects map knowledge more effectively. Keeping the transformer frozen loses 5%, showing the importance of fine-tuning when applying VLMs for specific tasks such as autonomous driving. Coarsening map resolution from 10 cm to 40 cm blurs lane edges and raises crash rate. Overall, these results demonstrate that early, token-level fusion of intent and map layout enables accurate, transparent, real-time driving.
Authors: Jobst Heitzig, Ram Potham
Abstract: Power is a key concept in AI safety: power-seeking as an instrumental goal, sudden or gradual disempowerment of humans, power balance in human-AI interaction and international AI governance. At the same time, power as the ability to pursue diverse goals is essential for wellbeing. This paper explores the idea of promoting both safety and wellbeing by forcing AI agents explicitly to empower humans and to manage the power balance between humans and AI agents in a desirable way. Using a principled, partially axiomatic approach, we design a parametrizable and decomposable objective function that represents an inequality- and risk-averse long-term aggregate of human power. It takes into account humans' bounded rationality and social norms, and, crucially, considers a wide variety of possible human goals. We derive algorithms for computing that metric by backward induction or approximating it via a form of multi-agent reinforcement learning from a given world model. We exemplify the consequences of (softly) maximizing this metric in a variety of paradigmatic situations and describe what instrumental sub-goals it will likely imply. Our cautious assessment is that softly maximizing suitable aggregate metrics of human power might constitute a beneficial objective for agentic AI systems that is safer than direct utility-based objectives.
Authors: Yihong Dong, Xue Jiang, Yongding Tao, Huanyu Liu, Kechi Zhang, Lili Mou, Rongyu Cao, Yingwei Ma, Jue Chen, Binhua Li, Zhi Jin, Fei Huang, Yongbin Li, Ge Li
Abstract: Reinforcement Learning with Verifiable Reward (RLVR) has significantly advanced the complex reasoning abilities of Large Language Models (LLMs). However, it struggles to break through the inherent capability boundaries of the base LLM, due to its essentially on-policy strategy coupled with LLM's immense action space and sparse reward. Critically, RLVR can lead to the capability boundary collapse, narrowing the LLM's problem-solving scope. To address this problem, we propose RL-PLUS, a novel hybrid-policy optimization approach for LLMs that synergizes internal exploitation with external data to achieve stronger reasoning capabilities and surpass the boundaries of base models. RL-PLUS integrates two core components, i.e., Multiple Importance Sampling to address distributional mismatch from external data, and Exploration-Based Advantage Function to guide the model towards high-value, unexplored reasoning paths. We provide both theoretical analysis and extensive experiments to demonstrate the superiority and generalizability of our approach. Compared with existing RLVR methods, RL-PLUS achieves 1) state-of-the-art performance on six math reasoning benchmarks; 2) superior performance on six out-of-distribution reasoning tasks; 3) consistent and significant gains across diverse model families, with average relative improvements up to 69.2\%. Moreover, the analysis of Pass@k curves indicates that RL-PLUS effectively resolves the capability boundary collapse problem.
Authors: Minghao Guo, Xi Zhu, Jingyuan Huang, Kai Mei, Yongfeng Zhang
Abstract: Graph Neural Networks (GNNs) have achieved remarkable success in graph-based learning by propagating information among neighbor nodes via predefined aggregation mechanisms. However, such fixed schemes often suffer from two key limitations. First, they cannot handle the imbalance in node informativeness -- some nodes are rich in information, while others remain sparse. Second, predefined message passing primarily leverages local structural similarity while ignoring global semantic relationships across the graph, limiting the model's ability to capture distant but relevant information. We propose Retrieval-augmented Graph Agentic Network (ReaGAN), an agent-based framework that empowers each node with autonomous, node-level decision-making. Each node acts as an agent that independently plans its next action based on its internal memory, enabling node-level planning and adaptive message propagation. Additionally, retrieval-augmented generation (RAG) allows nodes to access semantically relevant content and build global relationships in the graph. ReaGAN achieves competitive performance under few-shot in-context settings using a frozen LLM backbone without fine-tuning, showcasing the potential of agentic planning and local-global retrieval in graph learning.
Authors: Saba Ahmadi, Rabiul Awal, Ankur Sikarwar, Amirhossein Kazemnejad, Ge Ya Luo, Juan A. Rodriguez, Sai Rajeswar, Siva Reddy, Christopher Pal, Benno Krojer, Aishwarya Agrawal
Abstract: We explore three strategies to enhance performance on a wide range of image editing tasks: supervised fine-tuning (SFT), reinforcement learning (RL), and Chain-of-Thought (CoT) reasoning. In order to study all these components in one consistent framework, we adopt an autoregressive multimodal model that processes textual and visual tokens in a unified manner. We find RL combined with a large multi-modal LLM verifier to be the most effective of these strategies. As a result, we release EARL: Editing with Autoregression and RL, a strong RL-based image editing model that performs competitively on a diverse range of edits compared to strong baselines, despite using much less training data. Thus, EARL pushes the frontier of autoregressive multimodal models on image editing. We release our code, training data, and trained models at https://github.com/mair-lab/EARL.
Authors: Chengshuai Zhao, Zhen Tan, Pingchuan Ma, Dawei Li, Bohan Jiang, Yancheng Wang, Yingzhen Yang, Huan Liu
Abstract: Chain-of-Thought (CoT) prompting has been shown to improve Large Language Model (LLM) performance on various tasks. With this approach, LLMs appear to produce human-like reasoning steps before providing answers (a.k.a., CoT reasoning), which often leads to the perception that they engage in deliberate inferential processes. However, some initial findings suggest that CoT reasoning may be more superficial than it appears, motivating us to explore further. In this paper, we study CoT reasoning via a data distribution lens and investigate if CoT reasoning reflects a structured inductive bias learned from in-distribution data, allowing the model to conditionally generate reasoning paths that approximate those seen during training. Thus, its effectiveness is fundamentally bounded by the degree of distribution discrepancy between the training data and the test queries. With this lens, we dissect CoT reasoning via three dimensions: task, length, and format. To investigate each dimension, we design DataAlchemy, an isolated and controlled environment to train LLMs from scratch and systematically probe them under various distribution conditions. Our results reveal that CoT reasoning is a brittle mirage that vanishes when it is pushed beyond training distributions. This work offers a deeper understanding of why and when CoT reasoning fails, emphasizing the ongoing challenge of achieving genuine and generalizable reasoning.
Authors: Shitian Yang, Deyu Li, Xiaoke Jiang, Lei Zhang
Abstract: RGB-based 3D tasks, e.g., 3D detection, depth estimation, 3D keypoint estimation, still suffer from scarce, expensive annotations and a thin augmentation toolbox, since most image transforms, including resize and rotation, disrupt geometric consistency. In this paper, we introduce 3DRot, a plug-and-play augmentation that rotates and mirrors images about the camera's optical center while synchronously updating RGB images, camera intrinsics, object poses, and 3D annotations to preserve projective geometry-achieving geometry-consistent rotations and reflections without relying on any scene depth. We validate 3DRot with a classical 3D task, monocular 3D detection. On SUN RGB-D dataset, 3DRot raises $IoU_{3D}$ from 43.21 to 44.51, cuts rotation error (ROT) from 22.91$^\circ$ to 20.93$^\circ$, and boosts $mAP_{0.5}$ from 35.70 to 38.11. As a comparison, Cube R-CNN adds 3 other datasets together with SUN RGB-D for monocular 3D estimation, with a similar mechanism and test dataset, increases $IoU_{3D}$ from 36.2 to 37.8, boosts $mAP_{0.5}$ from 34.7 to 35.4. Because it operates purely through camera-space transforms, 3DRot is readily transferable to other 3D tasks.