Authors: Hiroshi Takahashi, Tomoharu Iwata, Atsutoshi Kumagai, Yuuki Yamanaka, Tomoya Yamashita
Abstract: Diffusion models are powerful generative models but often generate sensitive data that are unwanted by users, mainly because the unlabeled training data frequently contain such sensitive data. Since labeling all sensitive data in the large-scale unlabeled training data is impractical, we address this problem by using a small amount of labeled sensitive data. In this paper, we propose positive-unlabeled diffusion models, which prevent the generation of sensitive data using unlabeled and sensitive data. Our approach can approximate the evidence lower bound (ELBO) for normal (negative) data using only unlabeled and sensitive (positive) data. Therefore, even without labeled normal data, we can maximize the ELBO for normal data and minimize it for labeled sensitive data, ensuring the generation of only normal data. Through experiments across various datasets and settings, we demonstrated that our approach can prevent the generation of sensitive images without compromising image quality.
Authors: Qingyuan Jiang, Zhouyang Chi, Xiao Ma, Qirong Mao, Yang Yang, Jinhui Tang
Abstract: To address the modality learning degeneration caused by modality imbalance, existing multimodal learning~(MML) approaches primarily attempt to balance the optimization process of each modality from the perspective of model learning. However, almost all existing methods ignore the modality imbalance caused by unimodal data sampling, i.e., equal unimodal data sampling often results in discrepancies in informational content, leading to modality imbalance. Therefore, in this paper, we propose a novel MML approach called \underline{D}ata-aware \underline{U}nimodal \underline{S}ampling~(\method), which aims to dynamically alleviate the modality imbalance caused by sampling. Specifically, we first propose a novel cumulative modality discrepancy to monitor the multimodal learning process. Based on the learning status, we propose a heuristic and a reinforcement learning~(RL)-based data-aware unimodal sampling approaches to adaptively determine the quantity of sampled data at each iteration, thus alleviating the modality imbalance from the perspective of sampling. Meanwhile, our method can be seamlessly incorporated into almost all existing multimodal learning approaches as a plugin. Experiments demonstrate that \method~can achieve the best performance by comparing with diverse state-of-the-art~(SOTA) baselines.
Authors: Tianyi Huang
Abstract: Harmful Algal Blooms (HABs) pose severe threats to aquatic ecosystems and public health, resulting in substantial economic losses globally. Early detection is crucial but often hindered by the scarcity of high-quality datasets necessary for training reliable machine learning (ML) models. This study investigates the use of synthetic data augmentation using Gaussian Copulas to enhance ML-based HAB detection systems. Synthetic datasets of varying sizes (100-1,000 samples) were generated using relevant environmental features$\unicode{x2015}$water temperature, salinity, and UVB radiation$\unicode{x2015}$with corrected Chlorophyll-a concentration as the target variable. Experimental results demonstrate that moderate synthetic augmentation significantly improves model performance (RMSE reduced from 0.4706 to 0.1850; $p < 0.001$). However, excessive synthetic data introduces noise and reduces predictive accuracy, emphasizing the need for a balanced approach to data augmentation. These findings highlight the potential of synthetic data to enhance HAB monitoring systems, offering a scalable and cost-effective method for early detection and mitigation of ecological and public health risks.
Authors: Jianqi Yan (The University of Hong Kong), Alex P. Leung (The University of Hong Kong), Zhiyuan Pei (Macau University of Science and Technology), David C. Y. Hui (Chungnam National University), Sangin Kim (Chungnam National University)
Abstract: This work introduces a novel deep learning-based approach for gravitational wave anomaly detection, aiming to overcome the limitations of traditional matched filtering techniques in identifying unknown waveform gravitational wave signals. We introduce a modified convolutional neural network architecture inspired by ResNet that leverages residual blocks to extract high-dimensional features, effectively capturing subtle differences between background noise and gravitational wave signals. This network architecture learns a high-dimensional projection while preserving discrepancies with the original input, facilitating precise identification of gravitational wave signals. In our experiments, we implement an innovative data augmentation strategy that generates new data by computing the arithmetic mean of multiple signal samples while retaining the key features of the original signals. In the NSF HDR A3D3: Detecting Anomalous Gravitational Wave Signals competition, it is honorable for us (group name: easonyan123) to get to the first place at the end with our model achieving a true negative rate (TNR) of 0.9708 during development/validation phase and 0.9832 on an unseen challenge dataset during final/testing phase, the highest among all competitors. These results demonstrate that our method not only achieves excellent generalization performance but also maintains robust adaptability in addressing the complex uncertainties inherent in gravitational wave anomaly detection.
Authors: Fenglin Liu, Jinge Wu, Hongjian Zhou, Xiao Gu, Soheila Molaei, Anshul Thakur, Lei Clifton, Honghan Wu, David A. Clifton
Abstract: The application of Large Language Models (LLMs) to various clinical applications has attracted growing research attention. However, real-world clinical decision-making differs significantly from the standardized, exam-style scenarios commonly used in current efforts. In this paper, we present the RiskAgent system to perform a broad range of medical risk predictions, covering over 387 risk scenarios across diverse complex diseases, e.g., cardiovascular disease and cancer. RiskAgent is designed to collaborate with hundreds of clinical decision tools, i.e., risk calculators and scoring systems that are supported by evidence-based medicine. To evaluate our method, we have built the first benchmark MedRisk specialized for risk prediction, including 12,352 questions spanning 154 diseases, 86 symptoms, 50 specialties, and 24 organ systems. The results show that our RiskAgent, with 8 billion model parameters, achieves 76.33% accuracy, outperforming the most recent commercial LLMs, o1, o3-mini, and GPT-4.5, and doubling the 38.39% accuracy of GPT-4o. On rare diseases, e.g., Idiopathic Pulmonary Fibrosis (IPF), RiskAgent outperforms o1 and GPT-4.5 by 27.27% and 45.46% accuracy, respectively. Finally, we further conduct a generalization evaluation on an external evidence-based diagnosis benchmark and show that our RiskAgent achieves the best results. These encouraging results demonstrate the great potential of our solution for diverse diagnosis domains. To improve the adaptability of our model in different scenarios, we have built and open-sourced a family of models ranging from 1 billion to 70 billion parameters. Our code, data, and models are all available at https://github.com/AI-in-Health/RiskAgent.
Authors: Hetarth Chopra, Vidhi Rambhia, Vikram Adve
Abstract: As specialized large language models (LLMs) become increasingly prevalent, model merging methods are being used to combine them to create a single multi-task model without requiring any additional data or training. However, these approaches fall short when the objective of merging is to increase the downstream model's performance on a particular task-specific benchmark. In this work, we propose LEWIS (Layer Wise Sparsity), a guided model-merging framework that uses activation-based layer importance to dynamically adjust layer-wise task-vector sparsity required for the merge process. LEWIS uses a calibration dataset to prioritize critical layers during the task-vector pruning process required for model merging. This approach guides existing merging methods by preserving essential layer-wise task-specific knowledge while ensuring the merged model performs the best at benchmarks resembling the calibration dataset. Our experiments demonstrate the effectiveness of LEWIS with performance improvements of code instruction-following and math-solving models created through model merging up to 4 percent and 11.3 percent, respectively, outperforming unguided data-less model merging approaches that use uniform-sparsity.
Authors: Edoardo Zorzi, Alberto Castellini, Leonidas Bakopoulos, Georgios Chalkiadakis, Alessandro Farinelli
Abstract: Most offline RL algorithms return optimal policies but do not provide statistical guarantees on undesirable behaviors. This could generate reliability issues in safety-critical applications, such as in some multiagent domains where agents, and possibly humans, need to interact to reach their goals without harming each other. In this work, we propose a novel offline RL approach, inspired by Seldonian optimization, which returns policies with good performance and statistically guaranteed properties with respect to predefined undesirable behaviors. In particular, our focus is on Ad Hoc Teamwork settings, where agents must collaborate with new teammates without prior coordination. Our method requires only a pre-collected dataset, a set of candidate policies for our agent, and a specification about the possible policies followed by the other players -- it does not require further interactions, training, or assumptions on the type and architecture of the policies. We test our algorithm in Ad Hoc Teamwork problems and show that it consistently finds reliable policies while improving sample efficiency with respect to standard ML baselines.
Authors: Nikolaos Nakis, Chrysoula Kosma, Anastasia Brativnyk, Michail Chatzianastasis, Iakovos Evdaimon, Michalis Vazirgiannis
Abstract: Accurately predicting complex protein-protein interactions (PPIs) is crucial for decoding biological processes, from cellular functioning to disease mechanisms. However, experimental methods for determining PPIs are computationally expensive. Thus, attention has been recently drawn to machine learning approaches. Furthermore, insufficient effort has been made toward analyzing signed PPI networks, which capture both activating (positive) and inhibitory (negative) interactions. To accurately represent biological relationships, we present the Signed Two-Space Proximity Model (S2-SPM) for signed PPI networks, which explicitly incorporates both types of interactions, reflecting the complex regulatory mechanisms within biological systems. This is achieved by leveraging two independent latent spaces to differentiate between positive and negative interactions while representing protein similarity through proximity in these spaces. Our approach also enables the identification of archetypes representing extreme protein profiles. S2-SPM's superior performance in predicting the presence and sign of interactions in SPPI networks is demonstrated in link prediction tasks against relevant baseline methods. Additionally, the biological prevalence of the identified archetypes is confirmed by an enrichment analysis of Gene Ontology (GO) terms, which reveals that distinct biological tasks are associated with archetypal groups formed by both interactions. This study is also validated regarding statistical significance and sensitivity analysis, providing insights into the functional roles of different interaction types. Finally, the robustness and consistency of the extracted archetype structures are confirmed using the Bayesian Normalized Mutual Information (BNMI) metric, proving the model's reliability in capturing meaningful SPPI patterns.
Authors: Xiaochuan Gong, Jie Hao, Mingrui Liu
Abstract: Adam has become one of the most popular optimizers for training modern deep neural networks, such as transformers. However, its applicability is largely restricted to single-level optimization problems. In this paper, we aim to extend vanilla Adam to tackle bilevel optimization problems, which have important applications in machine learning, such as meta-learning. In particular, we study stochastic bilevel optimization problems where the lower-level function is strongly convex and the upper-level objective is nonconvex with potentially unbounded smoothness. This unbounded smooth objective function covers a broad class of neural networks, including transformers, which may exhibit non-Lipschitz gradients. In this work, we introduce AdamBO, a single-loop Adam-type method that achieves $\widetilde{O}(\epsilon^{-4})$ oracle complexity to find $\epsilon$-stationary points, where the oracle calls involve stochastic gradient or Hessian/Jacobian-vector product evaluations. The key to our analysis is a novel randomness decoupling lemma that provides refined control over the lower-level variable. We conduct extensive experiments on various machine learning tasks involving bilevel formulations with recurrent neural networks (RNNs) and transformers, demonstrating the effectiveness of our proposed Adam-type algorithm.
Authors: Jie Hao, Yuman Wu, Ali Payani, Myungjin Lee, Mingrui Liu
Abstract: We study the task of personalized federated fine-tuning with heterogeneous data in the context of language models, where clients collaboratively fine-tune a language model (e.g., BERT, GPT) without sharing their local data, achieving personalization simultaneously. While recent efforts have applied parameter-efficient fine-tuning techniques like low-rank adaptation (LoRA) in federated settings, they typically use single or multiple independent low-rank adapters with predefined maximal and minimal ranks, which may not be optimal for diverse data sources over clients. To address this issue, we propose PF2LoRA, a new personalized federated fine-tuning algorithm built on a novel \emph{automatic rank learning approach via two-level LoRA}. Given the pretrained language model whose weight is frozen, our algorithm aims to learn two levels of adaptation simultaneously: the first level aims to learn a common adapter for all clients, while the second level fosters individual client personalization. A key advantage of PF2LoRA is its ability to adaptively determine a suitable rank based on an individual client's data, rather than relying on a predefined rank that is agnostic to data heterogeneity. We present a synthetic example that highlights how PF2LoRA automatically learns the ground-truth rank for each client, tailoring the adaptation to match the properties of their individual data. Notably, this approach introduces minimal additional memory overhead, as the second-level adaptation comprises a small number of parameters compared to the first level. Our experiments on natural language understanding and generation tasks demonstrate that PF2LoRA significantly outperforms existing federated fine-tuning methods.
Authors: Abdullah Mamun, Asiful Arefeen, Susan B. Racette, Dorothy D. Sears, Corrie M. Whisner, Matthew P. Buman, Hassan Ghasemzadeh
Abstract: Postprandial hyperglycemia, marked by the blood glucose level exceeding the normal range after meals, is a critical indicator of progression toward type 2 diabetes in prediabetic and healthy individuals. A key metric for understanding blood glucose dynamics after eating is the postprandial area under the curve (PAUC). Predicting PAUC in advance based on a person's diet and activity level and explaining what affects postprandial blood glucose could allow an individual to adjust their lifestyle accordingly to maintain normal glucose levels. In this paper, we propose GlucoLens, an explainable machine learning approach to predict PAUC and hyperglycemia from diet, activity, and recent glucose patterns. We conducted a five-week user study with 10 full-time working individuals to develop and evaluate the computational model. Our machine learning model takes multimodal data including fasting glucose, recent glucose, recent activity, and macronutrient amounts, and provides an interpretable prediction of the postprandial glucose pattern. Our extensive analyses of the collected data revealed that the trained model achieves a normalized root mean squared error (NRMSE) of 0.123. On average, GlucoLense with a Random Forest backbone provides a 16% better result than the baseline models. Additionally, GlucoLens predicts hyperglycemia with an accuracy of 74% and recommends different options to help avoid hyperglycemia through diverse counterfactual explanations. Code available: https://github.com/ab9mamun/GlucoLens.
Authors: Zixuan Wang, Luis Zerkowski, Ilya Vidrin, Mariel Pettee
Abstract: Existing AI-generated dance methods primarily train on motion capture data from solo dance performances, but a critical feature of dance in nearly any genre is the interaction of two or more bodies in space. Moreover, many works at the intersection of AI and dance fail to incorporate the ideas and needs of the artists themselves into their development process, yielding models that produce far more useful insights for the AI community than for the dance community. This work addresses both needs of the field by proposing an AI method to model the complex interactions between pairs of dancers and detailing how the technical methodology can be shaped by ongoing co-creation with the artistic stakeholders who curated the movement data. Our model is a probability-and-attention-based Variational Autoencoder that generates a choreographic partner conditioned on an input dance sequence. We construct a custom loss function to enhance the smoothness and coherence of the generated choreography. Our code is open-source, and we also document strategies for other interdisciplinary research teams to facilitate collaboration and strong communication between artists and technologists.
Authors: William Merrill, Ashish Sabharwal
Abstract: Recent theoretical results show transformers cannot express sequential reasoning problems over long input lengths, intuitively because their computational depth is bounded. However, prior work treats the depth as a constant, leaving it unclear to what degree bounded depth may suffice for solving problems over short inputs, or how increasing the transformer's depth affects its expressive power. We address these questions by analyzing the expressive power of transformers whose depth can grow minimally with context length $n$. We show even highly uniform transformers with depth $\Theta(\log n)$ can express two important problems: recognizing regular languages, which captures state tracking abilities, and graph connectivity, which underlies multi-step reasoning. Notably, both of these problems cannot be expressed by fixed-depth transformers under standard complexity conjectures, demonstrating the expressivity benefit of growing depth. Moreover, our theory quantitatively predicts how depth must grow with input length to express these problems, showing that depth scaling is more efficient than scaling width or chain-of-thought steps. Empirically, we find our theoretical depth requirements for regular language recognition match the practical depth requirements of transformers remarkably well. Thus, our results clarify precisely how depth affects transformers' reasoning capabilities, providing potential practical insights for designing models that are better at sequential reasoning.
Authors: Dimitris G. Giovanis, Ellis Crabtree, Roger G. Ghanem, Ioannis G. kevrekidis
Abstract: A generative modeling framework is proposed that combines diffusion models and manifold learning to efficiently sample data densities on manifolds. The approach utilizes Diffusion Maps to uncover possible low-dimensional underlying (latent) spaces in the high-dimensional data (ambient) space. Two approaches for sampling from the latent data density are described. The first is a score-based diffusion model, which is trained to map a standard normal distribution to the latent data distribution using a neural network. The second one involves solving an It\^o stochastic differential equation in the latent space. Additional realizations of the data are generated by lifting the samples back to the ambient space using Double Diffusion Maps, a recently introduced technique typically employed in studying dynamical system reduction; here the focus lies in sampling densities rather than system dynamics. The proposed approaches enable sampling high dimensional data densities restricted to low-dimensional, a priori unknown manifolds. The efficacy of the proposed framework is demonstrated through a benchmark problem and a material with multiscale structure.
Authors: Chaitanya K. Joshi, Xiang Fu, Yi-Lun Liao, Vahe Gharakhanyan, Benjamin Kurt Miller, Anuroop Sriram, Zachary W. Ulissi
Abstract: Diffusion models are the standard toolkit for generative modelling of 3D atomic systems. However, for different types of atomic systems - such as molecules and materials - the generative processes are usually highly specific to the target system despite the underlying physics being the same. We introduce the All-atom Diffusion Transformer (ADiT), a unified latent diffusion framework for jointly generating both periodic materials and non-periodic molecular systems using the same model: (1) An autoencoder maps a unified, all-atom representations of molecules and materials to a shared latent embedding space; and (2) A diffusion model is trained to generate new latent embeddings that the autoencoder can decode to sample new molecules or materials. Experiments on QM9 and MP20 datasets demonstrate that jointly trained ADiT generates realistic and valid molecules as well as materials, exceeding state-of-the-art results from molecule and crystal-specific models. ADiT uses standard Transformers for both the autoencoder and diffusion model, resulting in significant speedups during training and inference compared to equivariant diffusion models. Scaling ADiT up to half a billion parameters predictably improves performance, representing a step towards broadly generalizable foundation models for generative chemistry. Open source code: https://github.com/facebookresearch/all-atom-diffusion-transformer
URLs: https://github.com/facebookresearch/all-atom-diffusion-transformer
Authors: Abdolvahhab Rostamijavanani, Shanwu Li, Yongchao Yang
Abstract: While linear systems have been useful in solving problems across different fields, the need for improved performance and efficiency has prompted them to operate in nonlinear modes. As a result, nonlinear models are now essential for the design and control of these systems. However, identifying a nonlinear system is more complicated than identifying a linear one. Therefore, modeling and identifying nonlinear systems are crucial for the design, manufacturing, and testing of complex systems. This study presents using advanced nonlinear methods based on deep learning for system identification. Two deep neural network models, LSTM autoencoder and Normalizing Flows, are explored for their potential to extract temporal features from time series data and relate them to system parameters, respectively. The presented framework offers a nonlinear approach to system identification, enabling it to handle complex systems. As case studies, we consider Duffing and Lorenz systems, as well as fluid flows such as flows over a cylinder and the 2-D lid-driven cavity problem. The results indicate that the presented framework is capable of capturing features and effectively relating them to system parameters, satisfying the identification requirements of nonlinear systems.
Authors: Sourabh Medapati, Priya Kasimbeg, Shankar Krishnan, Naman Agarwal, George Dahl
Abstract: If we want to train a neural network using any of the most popular optimization algorithms, we are immediately faced with a dilemma: how to set the various optimization and regularization hyperparameters? When computational resources are abundant, there are a variety of methods for finding good hyperparameter settings, but when resources are limited the only realistic choices are using standard default values of uncertain quality and provenance, or tuning only a couple of the most important hyperparameters via extremely limited handdesigned sweeps. Extending the idea of default settings to a modest tuning budget, Metz et al. (2020) proposed using ordered lists of well-performing hyperparameter settings, derived from a broad hyperparameter search on a large library of training workloads. However, to date, no practical and performant hyperparameter lists that generalize to representative deep learning workloads have been demonstrated. In this paper, we present hyperparameter lists for NAdamW derived from extensive experiments on the realistic workloads in the AlgoPerf: Training Algorithms benchmark. Our hyperparameter lists also include values for basic regularization techniques (i.e. weight decay, label smoothing, and dropout). In particular, our best NAdamW hyperparameter list performs well on AlgoPerf held-out workloads not used to construct it, and represents a compelling turn-key approach to tuning when restricted to five or fewer trials. It also outperforms basic learning rate/weight decay sweeps and an off-the-shelf Bayesian optimization tool when restricted to the same budget.
Authors: Sungwon Kim, Yoonho Lee, Yunhak Oh, Namkyeong Lee, Sukwon Yun, Junseok Lee, Sein Kim, Carl Yang, Chanyoung Park
Abstract: Federated Learning (FL) on graphs enables collaborative model training to enhance performance without compromising the privacy of each client. However, existing methods often overlook the mutable nature of graph data, which frequently introduces new nodes and leads to shifts in label distribution. Since they focus solely on performing well on each client's local data, they are prone to overfitting to their local distributions (i.e., local overfitting), which hinders their ability to generalize to unseen data with diverse label distributions. In contrast, our proposed method, FedLoG, effectively tackles this issue by mitigating local overfitting. Our model generates global synthetic data by condensing the reliable information from each class representation and its structural information across clients. Using these synthetic data as a training set, we alleviate the local overfitting problem by adaptively generalizing the absent knowledge within each local dataset. This enhances the generalization capabilities of local models, enabling them to handle unseen data effectively. Our model outperforms baselines in our proposed experimental settings, which are designed to measure generalization power to unseen data in practical scenarios. Our code is available at https://github.com/sung-won-kim/FedLoG
Authors: Aleksandrs Slivkins, Yunzong Xu, Shiliang Zuo
Abstract: We study the greedy (exploitation-only) algorithm in bandit problems with a known reward structure. We allow arbitrary finite reward structures, while prior work focused on a few specific ones. We fully characterize when the greedy algorithm asymptotically succeeds or fails, in the sense of sublinear vs. linear regret as a function of time. Our characterization identifies a partial identifiability property of the problem instance as the necessary and sufficient condition for the asymptotic success. Notably, once this property holds, the problem becomes easy -- any algorithm will succeed (in the same sense as above), provided it satisfies a mild non-degeneracy condition. We further extend our characterization to contextual bandits and interactive decision-making with arbitrary feedback, and demonstrate its broad applicability across various examples.
Authors: Zhipeng Zhou, Ziqiao Meng, Pengcheng Wu, Peilin Zhao, Chunyan Miao
Abstract: Multi-task learning (MTL) is a widely explored paradigm that enables the simultaneous learning of multiple tasks using a single model. Despite numerous solutions, the key issues of optimization conflict and task imbalance remain under-addressed, limiting performance. Unlike existing optimization-based approaches that typically reweight task losses or gradients to mitigate conflicts or promote progress, we propose a novel approach based on Continual Optimization with Symmetry Teleportation (COST). During MTL optimization, when an optimization conflict arises, we seek an alternative loss-equivalent point on the loss landscape to reduce conflict. Specifically, we utilize a low-rank adapter (LoRA) to facilitate this practical teleportation by designing convergent, loss-invariant objectives. Additionally, we introduce a historical trajectory reuse strategy to continually leverage the benefits of advanced optimizers. Extensive experiments on multiple mainstream datasets demonstrate the effectiveness of our approach. COST is a plug-and-play solution that enhances a wide range of existing MTL methods. When integrated with state-of-the-art methods, COST achieves superior performance.
Authors: Chengxin Zhang, Yujie Liu, Quan Chen
Abstract: As the demand for computational power increases, high-bandwidth memory (HBM) has become a critical technology for next-generation computing systems. However, the widespread adoption of HBM presents significant thermal management challenges, particularly in multilayer through-silicon-via (TSV) stacked structures under varying thermal conditions, where accurate prediction of junction temperature and hotspot position is essential during the early design. This work develops a data-driven neural network model for the fast prediction of junction temperature and hotspot position in 3D HBM chiplets. The model, trained with a data set of $13,494$ different combinations of thermal condition parameters, sampled from a vast parameter space characterized by high-dimensional combination (up to $3^{27}$), can accurately and quickly infer the junction temperature and hotspot position for any thermal conditions in the parameter space. Moreover, it shows good generalizability for other thermal conditions not considered in the parameter space. The data set is constructed using accurate finite element solvers. This method not only minimizes the reliance on costly experimental tests and extensive computational resources for finite element analysis but also accelerates the design and optimization of complex HBM systems, making it a valuable tool for improving thermal management and performance in high-performance computing applications.
Authors: Ziruo Hao, Zhenhua Cui, Tao Yang, Bo Hu, Xiaofeng Wu, Hui Feng
Abstract: Federated learning (FL) has provided a new methodology for coordinating a group of clients to train a machine learning model collaboratively, bringing an efficient paradigm in edge intelligence. Despite its promise, FL faces several critical challenges in practical applications involving edge devices, such as data heterogeneity and delays stemming from communication and computation constraints. This paper examines the impact of unknown causes of delay on training performance in an Asynchronous Federated Learning (AFL) system with data heterogeneity. Initially, an asynchronous error definition is proposed, based on which the solely adverse impact of data heterogeneity is theoretically analyzed within the traditional Synchronous Federated Learning (SFL) framework. Furthermore, Asynchronous Updates with Delayed Gradients (AUDG), a conventional AFL scheme, is discussed. Investigation into AUDG reveals that the negative influence of data heterogeneity is correlated with delays, while a shorter average delay from a specific client does not consistently enhance training performance. In order to compensate for the scenarios where AUDG are not adapted, Pseudo-synchronous Updates by Reusing Delayed Gradients (PSURDG) is proposed, and its theoretical convergence is analyzed. In both AUDG and PSURDG, only a random set of clients successfully transmits their updated results to the central server in each iteration. The critical difference between them lies in whether the delayed information is reused. Finally, both schemes are validated and compared through theoretical analysis and simulations, demonstrating more intuitively that discarding outdated information due to time delays is not always the best approach.
Authors: Shahrzad Kiani, Franziska Boenisch, Stark C. Draper
Abstract: Federated Learning (FL) is the standard protocol for collaborative learning. In FL, multiple workers jointly train a shared model. They exchange model updates calculated on their data, while keeping the raw data itself local. Since workers naturally form groups based on common interests and privacy policies, we are motivated to extend standard FL to reflect a setting with multiple, potentially overlapping groups. In this setup where workers can belong and contribute to more than one group at a time, complexities arise in understanding privacy leakage and in adhering to privacy policies. To address the challenges, we propose differential private overlapping grouped learning (DPOGL), a novel method to implement privacy guarantees within overlapping groups. Under the honest-but-curious threat model, we derive novel privacy guarantees between arbitrary pairs of workers. These privacy guarantees describe and quantify two key effects of privacy leakage in DP-OGL: propagation delay, i.e., the fact that information from one group will leak to other groups only with temporal offset through the common workers and information degradation, i.e., the fact that noise addition over model updates limits information leakage between workers. Our experiments show that applying DP-OGL enhances utility while maintaining strong privacy compared to standard FL setups.
Authors: Hongyao Tang
Abstract: Learning the optimal policy from a random network initialization is the theme of deep Reinforcement Learning (RL). As the scale of DRL training increases, treating DRL policy network weights as a new data modality and exploring the potential becomes appealing and possible. In this work, we focus on the policy learning path in deep RL, represented by the trajectory of network weights of historical policies, which reflects the evolvement of the policy learning process. Taking the idea of trajectory modeling with Transformer, we propose Transformer as Implicit Policy Learner (TIPL), which processes policy network weights in an autoregressive manner. We collect the policy learning path data by running independent RL training trials, with which we then train our TIPL model. In the experiments, we demonstrate that TIPL is able to fit the implicit dynamics of policy learning and perform the optimization of policy network by inference.
Authors: Yuqing Wang, Xiao Yang
Abstract: With the rapid expansion of cloud computing infrastructure, energy consumption has become a critical challenge, driving the need for accurate and efficient prediction models. This study proposes a novel Vector Weighted Average Kernel Extreme Learning Machine (VWAA-KELM) model to enhance energy consumption prediction in cloud computing environments. By integrating a vector weighted average algorithm (VWAA) with kernel extreme learning machine (KELM), the proposed model dynamically adjusts feature weights and optimizes kernel functions, significantly improving prediction accuracy and generalization. Experimental results demonstrate the superior performance of VWAA-KELM: 94.7% of test set prediction errors fall within [0, 50] units, with only three cases exceeding 100 units, indicating strong stability. The model achieves a coefficient of determination (R2) of 0.987 in the training set (RMSE = 28.108, RPD = 8.872) and maintains excellent generalization with R2 = 0.973 in the test set (RMSE = 43.227, RPD = 6.202). Visual analysis confirms that predicted values closely align with actual energy consumption trends, avoiding overfitting while capturing nonlinear dependencies. A key innovation of this study is the introduction of adaptive feature weighting, allowing the model to dynamically assign importance to different input parameters, thereby enhancing high-dimensional data processing. This advancement provides a scalable and efficient approach for optimizing cloud data center energy consumption. Beyond cloud computing, the proposed hybrid framework has broader applications in Internet of Things (IoT) and edge computing, supporting real-time energy management and intelligent resource allocation.
Authors: Seth Karten, Andy Luu Nguyen, Chi Jin
Abstract: We introduce Pok\'eChamp, a minimax agent powered by Large Language Models (LLMs) for Pok\'emon battles. Built on a general framework for two-player competitive games, Pok\'eChamp leverages the generalist capabilities of LLMs to enhance minimax tree search. Specifically, LLMs replace three key modules: (1) player action sampling, (2) opponent modeling, and (3) value function estimation, enabling the agent to effectively utilize gameplay history and human knowledge to reduce the search space and address partial observability. Notably, our framework requires no additional LLM training. We evaluate Pok\'eChamp in the popular Gen 9 OU format. When powered by GPT-4o, it achieves a win rate of 76% against the best existing LLM-based bot and 84% against the strongest rule-based bot, demonstrating its superior performance. Even with an open-source 8-billion-parameter Llama 3.1 model, Pok\'eChamp consistently outperforms the previous best LLM-based bot, Pok\'ellmon powered by GPT-4o, with a 64% win rate. Pok\'eChamp attains a projected Elo of 1300-1500 on the Pok\'emon Showdown online ladder, placing it among the top 30%-10% of human players. In addition, this work compiles the largest real-player Pok\'emon battle dataset, featuring over 3 million games, including more than 500k high-Elo matches. Based on this dataset, we establish a series of battle benchmarks and puzzles to evaluate specific battling skills. We further provide key updates to the local game engine. We hope this work fosters further research that leverage Pok\'emon battle as benchmark to integrate LLM technologies with game-theoretic algorithms addressing general multiagent problems. Videos, code, and dataset available at https://sites.google.com/view/pokechamp-llm.
Authors: Lijia Yu, Yibo Miao, Yifan Zhu, Xiao-Shan Gao, Lijun Zhang
Abstract: The primary objective of learning methods is generalization. Classic uniform generalization bounds, which rely on VC-dimension or Rademacher complexity, fail to explain the significant attribute that over-parameterized models in deep learning exhibit nice generalizability. On the other hand, algorithm-dependent generalization bounds, like stability bounds, often rely on strict assumptions. To establish generalizability under less stringent assumptions, this paper investigates the generalizability of neural networks that minimize or approximately minimize empirical risk. We establish a lower bound for population accuracy based on the expressiveness of these networks, which indicates that with an adequate large number of training samples and network sizes, these networks, including over-parameterized ones, can generalize effectively. Additionally, we provide a necessary condition for generalization, demonstrating that, for certain data distributions, the quantity of training data required to ensure generalization exceeds the network size needed to represent the corresponding data distribution. Finally, we provide theoretical insights into several phenomena in deep learning, including robust generalization, importance of over-parameterization, and effect of loss function on generalization.
Authors: Congxi Xiao, Jingbo Zhou, Yixiong Xiao, Xinjiang Lu, Le Zhang, Hui Xiong
Abstract: We present TimeFound, an encoder-decoder transformer-based time series foundation model for out-of-the-box zero-shot forecasting. To handle time series data from various domains, TimeFound employs a multi-resolution patching strategy to capture complex temporal patterns at multiple scales. We pre-train our model with two sizes (200M and 710M parameters) on a large time-series corpus comprising both real-world and synthetic datasets. Over a collection of unseen datasets across diverse domains and forecasting horizons, our empirical evaluations suggest that TimeFound can achieve superior or competitive zero-shot forecasting performance, compared to state-of-the-art time series foundation models.
Authors: Bruce Nguyen, Yan Zhang
Abstract: Diabetes is a prevalent chronic disease with significant health and economic burdens worldwide. Early prediction and diagnosis can aid in effective management and prevention of complications. This study explores the use of machine learning models to predict diabetes based on lifestyle factors using data from the Behavioral Risk Factor Surveillance System (BRFSS) 2015 survey. The dataset consists of 21 lifestyle and health-related features, capturing aspects such as physical activity, diet, mental health, and socioeconomic status. Three classification models, Decision Tree, K-Nearest Neighbors (KNN), and Logistic Regression, are implemented and evaluated to determine their predictive performance. The models are trained and tested using a balanced dataset, and their performances are assessed based on accuracy, precision, recall, and F1-score. The results indicate that the Decision Tree, KNN, and Logistic Regression achieve an accuracy of 0.74, 0.72, and 0.75, respectively, with varying strengths in precision and recall. The findings highlight the potential of machine learning in diabetes prediction and suggest future improvements through feature selection and ensemble learning techniques.
Authors: Kaiwen Wu, Craig Sanders, Benjamin Letham, Phillip Guan
Abstract: Gaussian processes (GPs) are powerful models for human-in-the-loop experiments due to their flexibility and well-calibrated uncertainty. However, GPs modeling human responses typically ignore auxiliary information, including a priori domain expertise and non-task performance information like user confidence ratings. We propose mixed likelihood variational GPs to leverage auxiliary information, which combine multiple likelihoods in a single evidence lower bound to model multiple types of data. We demonstrate the benefits of mixing likelihoods in three real-world experiments with human participants. First, we use mixed likelihood training to impose prior knowledge constraints in GP classifiers, which accelerates active learning in a visual perception task where users are asked to identify geometric errors resulting from camera position errors in virtual reality. Second, we show that leveraging Likert scale confidence ratings by mixed likelihood training improves model fitting for haptic perception of surface roughness. Lastly, we show that Likert scale confidence ratings improve human preference learning in robot gait optimization. The modeling performance improvements found using our framework across this diverse set of applications illustrates the benefits of incorporating auxiliary information into active learning and preference learning by using mixed likelihoods to jointly model multiple inputs.
Authors: Fengchen Gu, Zhengyong Jiang, \'Angel F. Garc\'ia-Fern\'andez, Angelos Stefanidis, Jionglong Su, Huakang Li
Abstract: Portfolio management remains a crucial challenge in finance, with traditional methods often falling short in complex and volatile market environments. While deep reinforcement approaches have shown promise, they still face limitations in dynamic risk management, exploitation of temporal markets, and incorporation of complex trading strategies such as short-selling. These limitations can lead to suboptimal portfolio performance, increased vulnerability to market volatility, and missed opportunities in capturing potential returns from diverse market conditions. This paper introduces a Deep Reinforcement Learning Portfolio Management Framework with Time-Awareness and Short-Selling (MTS), offering a robust and adaptive strategy for sustainable investment performance. This framework utilizes a novel encoder-attention mechanism to address the limitations by incorporating temporal market characteristics, a parallel strategy for automated short-selling based on market trends, and risk management through innovative Incremental Conditional Value at Risk, enhancing adaptability and performance. Experimental validation on five diverse datasets from 2019 to 2023 demonstrates MTS's superiority over traditional algorithms and advanced machine learning techniques. MTS consistently achieves higher cumulative returns, Sharpe, Omega, and Sortino ratios, underscoring its effectiveness in balancing risk and return while adapting to market dynamics. MTS demonstrates an average relative increase of 30.67% in cumulative returns and 29.33% in Sharpe ratio compared to the next best-performing strategies across various datasets.
Authors: Varatheepan Paramanayakam, Andreas Karatzas, Dimitrios Stamoulis, Iraklis Anagnostopoulos
Abstract: Edge computing systems struggle to efficiently manage multiple concurrent deep neural network (DNN) workloads while meeting strict latency requirements, minimizing power consumption, and maintaining environmental sustainability. This paper introduces Ecomap, a sustainability-driven framework that dynamically adjusts the maximum power threshold of edge devices based on real-time carbon intensity. Ecomap incorporates the innovative use of mixed-quality models, allowing it to dynamically replace computationally heavy DNNs with lighter alternatives when latency constraints are violated, ensuring service responsiveness with minimal accuracy loss. Additionally, it employs a transformer-based estimator to guide efficient workload mappings. Experimental results using NVIDIA Jetson AGX Xavier demonstrate that Ecomap reduces carbon emissions by an average of 30% and achieves a 25% lower carbon delay product (CDP) compared to state-of-the-art methods, while maintaining comparable or better latency and power efficiency.
Authors: Manh Cuong Dao, Phi Le Nguyen, Thao Nguyen Truong, Trong Nghia Hoang
Abstract: Offline optimization is an important task in numerous material engineering domains where online experimentation to collect data is too expensive and needs to be replaced by an in silico maximization of a surrogate of the black-box function. Although such a surrogate can be learned from offline data, its prediction might not be reliable outside the offline data regime, which happens when the surrogate has narrow prediction margin and is (therefore) sensitive to small perturbations of its parameterization. This raises the following questions: (1) how to regulate the sensitivity of a surrogate model; and (2) whether conditioning an offline optimizer with such less sensitive surrogate will lead to better optimization performance. To address these questions, we develop an optimizable sensitivity measurement for the surrogate model, which then inspires a sensitivity-informed regularizer that is applicable to a wide range of offline optimizers. This development is both orthogonal and synergistic to prior research on offline optimization, which is demonstrated in our extensive experiment benchmark.
Authors: Sicong Liu, Bin Guo, Shiyan Luo, Yuzhan Wang, Hao Luo, Cheng Fang, Yuan Xu, Ke Ma, Yao Li, Zhiwen Yu
Abstract: There are many deep learning (DL) powered mobile and wearable applications today continuously and unobtrusively sensing the ambient surroundings to enhance all aspects of human lives.To enable robust and private mobile sensing, DL models are often deployed locally on resource-constrained mobile devices using techniques such as model compression or offloading.However, existing methods, either front-end algorithm level (i.e. DL model compression/partitioning) or back-end scheduling level (i.e. operator/resource scheduling), cannot be locally online because they require offline retraining to ensure accuracy or rely on manually pre-defined strategies, struggle with dynamic adaptability.The primary challenge lies in feeding back runtime performance from the back-end level to the front-end level optimization decision. Moreover, the adaptive mobile DL model porting middleware with cross-level co-adaptation is less explored, particularly in mobile environments with diversity and dynamics. In response, we introduce CrowdHMTware, a dynamic context-adaptive DL model deployment middleware for heterogeneous mobile devices. It establishes an automated adaptation loop between cross-level functional components, i.e. elastic inference, scalable offloading, and model-adaptive engine, enhancing scalability and adaptability. Experiments with four typical tasks across 15 platforms and a real-world case study demonstrate that CrowdHMTware can effectively scale DL model, offloading, and engine actions across diverse platforms and tasks. It hides run-time system issues from developers, reducing the required developer expertise.
Authors: Arsenii Mustafin, Aleksei Pakharev, Alex Olshevsky, Ioannis Ch. Paschalidis
Abstract: We build on a recently introduced geometric interpretation of Markov Decision Processes (MDPs) to analyze classical MDP-solving algorithms: Value Iteration (VI) and Policy Iteration (PI). First, we develop a geometry-based analytical apparatus, including a transformation that modifies the discount factor $\gamma$, to improve convergence guarantees for these algorithms in several settings. In particular, one of our results identifies a rotation component in the VI method, and as a consequence shows that when a Markov Reward Process (MRP) induced by the optimal policy is irreducible and aperiodic, the asymptotic convergence rate of value iteration is strictly smaller than $\gamma$.
Authors: Alireza Habibi, Saeed Ghoorchian, Setareh Maghsudi
Abstract: The complexity of online decision-making under uncertainty stems from the requirement of finding a balance between exploiting known strategies and exploring new possibilities. Naturally, the uncertainty type plays a crucial role in developing decision-making strategies that manage complexity effectively. In this paper, we focus on a specific form of uncertainty known as epistemic ambivalence (EA), which emerges from conflicting pieces of evidence or contradictory experiences. It creates a delicate interplay between uncertainty and confidence, distinguishing it from epistemic uncertainty that typically diminishes with new information. Indeed, ambivalence can persist even after additional knowledge is acquired. To address this phenomenon, we propose a novel framework, called the epistemically ambivalent Markov decision process (EA-MDP), aiming to understand and control EA in decision-making processes. This framework incorporates the concept of a quantum state from the quantum mechanics formalism, and its core is to assess the probability and reward of every possible outcome. We calculate the reward function using quantum measurement techniques and prove the existence of an optimal policy and an optimal value function in the EA-MDP framework. We also propose the EA-epsilon-greedy Q-learning algorithm. To evaluate the impact of EA on decision-making and the expedience of our framework, we study two distinct experimental setups, namely the two-state problem and the lattice problem. Our results show that using our methods, the agent converges to the optimal policy in the presence of EA.
Authors: Maciej Krzysztof Zuziak, Roberto Pellungrini, Salvatore Rinzivillo
Abstract: Federated Learning (FL) is a widespread and well adopted paradigm of decentralized learning that allows training one model from multiple sources without the need to directly transfer data between participating clients. Since its inception in 2015, it has been divided into numerous sub-fields that deal with application-specific issues, be it data heterogeneity or resource allocation. One such sub-field, Clustered Federated Learning (CFL), is dealing with the problem of clustering the population of clients into separate cohorts to deliver personalized models. Although few remarkable works have been published in this domain, the problem is still largely unexplored, as its basic assumption and settings are slightly different from standard FL. In this work, we present One-Shot Clustered Federated Learning (OCFL), a clustering-agnostic algorithm that can automatically detect the earliest suitable moment for clustering. Our algorithm is based on the computation of cosine similarity between gradients of the clients and a temperature measure that detects when the federated model starts to converge. We empirically evaluate our methodology by testing various one-shot clustering algorithms for over thirty different tasks on three benchmark datasets. Our experiments showcase the good performance of our approach when used to perform CFL in an automated manner without the need to adjust hyperparameters.
Authors: Manh Cuong Dao, Phi Le Nguyen, Thao Nguyen Truong, Trong Nghia Hoang
Abstract: Offline optimization has recently emerged as an increasingly popular approach to mitigate the prohibitively expensive cost of online experimentation. The key idea is to learn a surrogate of the black-box function that underlines the target experiment using a static (offline) dataset of its previous input-output queries. Such an approach is, however, fraught with an out-of-distribution issue where the learned surrogate becomes inaccurate outside the offline data regimes. To mitigate this, existing offline optimizers have proposed numerous conditioning techniques to prevent the learned surrogate from being too erratic. Nonetheless, such conditioning strategies are often specific to particular surrogate or search models, which might not generalize to a different model choice. This motivates us to develop a model-agnostic approach instead, which incorporates a notion of model sharpness into the training loss of the surrogate as a regularizer. Our approach is supported by a new theoretical analysis demonstrating that reducing surrogate sharpness on the offline dataset provably reduces its generalized sharpness on unseen data. Our analysis extends existing theories from bounding generalized prediction loss (on unseen data) with loss sharpness to bounding the worst-case generalized surrogate sharpness with its empirical estimate on training data, providing a new perspective on sharpness regularization. Our extensive experimentation on a diverse range of optimization tasks also shows that reducing surrogate sharpness often leads to significant improvement, marking (up to) a noticeable 9.6% performance boost. Our code is publicly available at https://github.com/cuong-dm/IGNITE
Authors: Junhao Shi, Qinyuan Cheng, Zhaoye Fei, Yining Zheng, Qipeng Guo, Xipeng Qiu
Abstract: Aligning powerful AI models on tasks that surpass human evaluation capabilities is the central problem of \textbf{superalignment}. To address this problem, weak-to-strong generalization aims to elicit the capabilities of strong models through weak supervisors and ensure that the behavior of strong models aligns with the intentions of weak supervisors without unsafe behaviors such as deception. Although weak-to-strong generalization exhibiting certain generalization capabilities, strong models exhibit significant overfitting in weak-to-strong generalization: Due to the strong fit ability of strong models, erroneous labels from weak supervisors may lead to overfitting in strong models. In addition, simply filtering out incorrect labels may lead to a degeneration in question quality, resulting in a weak generalization ability of strong models on hard questions. To mitigate overfitting in weak-to-strong generalization, we propose a two-stage framework that simultaneously improves the quality of supervision signals and the quality of input questions. Experimental results in three series of large language models and two mathematical benchmarks demonstrate that our framework significantly improves PGR compared to naive weak-to-strong generalization, even achieving up to 100\% PGR on some models.
Authors: Yixiang Sun, Haotian Fu, Michael Littman, George Konidaris
Abstract: We propose DRAGO, a novel approach for continual model-based reinforcement learning aimed at improving the incremental development of world models across a sequence of tasks that differ in their reward functions but not the state space or dynamics. DRAGO comprises two key components: Synthetic Experience Rehearsal, which leverages generative models to create synthetic experiences from past tasks, allowing the agent to reinforce previously learned dynamics without storing data, and Regaining Memories Through Exploration, which introduces an intrinsic reward mechanism to guide the agent toward revisiting relevant states from prior tasks. Together, these components enable the agent to maintain a comprehensive and continually developing world model, facilitating more effective learning and adaptation across diverse environments. Empirical evaluations demonstrate that DRAGO is able to preserve knowledge across tasks, achieving superior performance in various continual learning scenarios.
Authors: Nadav Dym, Jianfeng Lu, Matan Mizrachi
Abstract: Motivated by applications for simulating quantum many body functions, we propose a new universal ansatz for approximating anti-symmetric functions. The main advantage of this ansatz over previous alternatives is that it is bi-Lipschitz with respect to a naturally defined metric. As a result, we are able to obtain quantitative approximation results for approximation of Lipschitz continuous antisymmetric functions. Moreover, we provide preliminary experimental evidence to the improved performance of this ansatz for learning antisymmetric functions.
Authors: Inkyu Kim, Sangkeum Lee, Haechan Jeong, Sarvar Hussain Nengroo, Dongsoo Har
Abstract: Satellite communication systems (SCSs) used for tactical purposes require robust security and anti-jamming capabilities, making frequency hopping (FH) a powerful option. However, the current FH systems face challenges due to significant interference from other devices and the considerable path loss inherent in satellite communication. This misalignment leads to inefficient synchronization, crucial for maintaining reliable communication. Traditional methods, such as those employing long short-term memory (LSTM) networks, have made improvements, but they still struggle in dynamic conditions of satellite environments. This paper presents a novel method for synchronizing FH signals in tactical SCSs by combining serial search and reinforcement learning to achieve coarse and fine acquisition, respectively. The mathematical analysis and simulation results demonstrate that the proposed method reduces the average number of hops required for synchronization by 58.17% and mean squared error (MSE) of the uplink hop timing estimation by 76.95%, as compared to the conventional serial search method. Comparing with the early late gate synchronization method based on serial search and use of LSTM network, the average number of hops for synchronization is reduced by 12.24% and the MSE by 18.5%.
Authors: Giovanni Di Gennaro, Amedeo Buonanno, Gianmarco Romano, Stefano Buzzi, Francesco A. N. Palmieri
Abstract: This study addresses the challenge of access point (AP) and user equipment (UE) association in cell-free massive MIMO networks. It introduces a deep learning algorithm leveraging Bidirectional Long Short-Term Memory cells and a hybrid probabilistic methodology for weight updating. This approach enhances scalability by adapting to variations in the number of UEs without requiring retraining. Additionally, the study presents a training methodology that improves scalability not only with respect to the number of UEs but also to the number of APs. Furthermore, a variant of the proposed AP-UE algorithm ensures robustness against pilot contamination effects, a critical issue arising from pilot reuse in channel estimation. Extensive numerical results validate the effectiveness and adaptability of the proposed methods, demonstrating their superiority over widely used heuristic alternatives.
Authors: Fabio Michele Russo, Carlo Metta, Anna Monreale, Salvatore Rinzivillo, Fabio Pinelli
Abstract: As predictive machine learning models become increasingly adopted and advanced, their role has evolved from merely predicting outcomes to actively shaping them. This evolution has underscored the importance of Trustworthy AI, highlighting the necessity to extend our focus beyond mere accuracy and toward a comprehensive understanding of these models' behaviors within the specific contexts of their applications. To further progress in explainability, we introduce Poem, Prefetched Offline Explanation Model, a model-agnostic, local explainability algorithm for image data. The algorithm generates exemplars, counterexemplars and saliency maps to provide quick and effective explanations suitable for time-sensitive scenarios. Leveraging an existing local algorithm, \poem{} infers factual and counterfactual rules from data to create illustrative examples and opposite scenarios with an enhanced stability by design. A novel mechanism then matches incoming test points with an explanation base and produces diverse exemplars, informative saliency maps and believable counterexemplars. Experimental results indicate that Poem outperforms its predecessor Abele in speed and ability to generate more nuanced and varied exemplars alongside more insightful saliency maps and valuable counterexemplars.
Authors: Shuang Liu, Yihan Wang, Yifan Zhu, Yibo Miao, Xiao-Shan Gao
Abstract: Wasserstein distributionally robust optimization (WDRO) optimizes against worst-case distributional shifts within a specified uncertainty set, leading to enhanced generalization on unseen adversarial examples, compared to standard adversarial training which focuses on pointwise adversarial perturbations. However, WDRO still suffers fundamentally from the robust overfitting problem, as it does not consider statistical error. We address this gap by proposing a novel robust optimization framework under a new uncertainty set for adversarial noise via Wasserstein distance and statistical error via Kullback-Leibler divergence, called the Statistically Robust WDRO. We establish a robust generalization bound for the new optimization framework, implying that out-of-distribution adversarial performance is at least as good as the statistically robust training loss with high probability. Furthermore, we derive conditions under which Stackelberg and Nash equilibria exist between the learner and the adversary, giving an optimal robust model in certain sense. Finally, through extensive experiments, we demonstrate that our method significantly mitigates robust overfitting and enhances robustness within the framework of WDRO.
Authors: Tim Maurer, Abdulrahman Mohamed Selim, Hasan Md Tusfiqur Alam, Matthias Eiletz, Michael Barz, Daniel Sonntag
Abstract: This paper presents InFL-UX, an interactive, proof-of-concept browser-based Federated Learning (FL) toolkit designed to integrate user contributions seamlessly into the machine learning (ML) workflow. InFL-UX enables users across multiple devices to upload datasets, define classes, and collaboratively train classification models directly in the browser using modern web technologies. Unlike traditional FL toolkits, which often focus on backend simulations, InFL-UX provides a simple user interface for researchers to explore how users interact with and contribute to FL systems in real-world, interactive settings. By prioritising usability and decentralised model training, InFL-UX bridges the gap between FL and Interactive Machine Learning (IML), empowering non-technical users to actively participate in ML classification tasks.
Authors: Izzy Newsham, Luka Kova\v{c}evi\'c, Richard Moulange, Nan Rosemary Ke, Sach Mukherjee
Abstract: Genes, proteins and other biological entities influence one another via causal molecular networks. Causal relationships in such networks are mediated by complex and diverse mechanisms, through latent variables, and are often specific to cellular context. It remains challenging to characterise such networks in practice. Here, we present a novel framework to evaluate large language models (LLMs) for zero-shot inference of causal relationships in biology. In particular, we systematically evaluate causal claims obtained from an LLM using real-world interventional data. This is done over one hundred variables and thousands of causal hypotheses. Furthermore, we consider several prompting and retrieval-augmentation strategies, including large, and potentially conflicting, collections of scientific articles. Our results show that with tailored augmentation and prompting, even relatively small LLMs can capture meaningful aspects of causal structure in biological systems. This supports the notion that LLMs could act as orchestration tools in biological discovery, by helping to distil current knowledge in ways amenable to downstream analysis. Our approach to assessing LLMs with respect to experimental data is relevant for a broad range of problems at the intersection of causal learning, LLMs and scientific discovery.
Authors: Joana Sim\~oes, Jo\~ao Correia
Abstract: Automated Machine Learning (AutoML) gained popularity due to the increased demand for Machine Learning (ML) specialists, allowing them to apply ML techniques effortlessly and quickly. AutoML implementations use optimisation methods to identify the most effective ML solution for a given dataset, aiming to improve one or more predefined metrics. However, most implementations focus on model selection and hyperparameter tuning. Despite being an important factor in obtaining high-performance ML systems, data quality is usually an overlooked part of AutoML and continues to be a manual and time-consuming task. This work presents EDCA, an Evolutionary Data Centric AutoML framework. In addition to the traditional tasks such as selecting the best models and hyperparameters, EDCA enhances the given data by optimising data processing tasks such as data reduction and cleaning according to the problems' needs. All these steps create an ML pipeline that is optimised by an evolutionary algorithm. To assess its effectiveness, EDCA was compared to FLAML and TPOT, two frameworks at the top of the AutoML benchmarks. The frameworks were evaluated in the same conditions using datasets from AMLB classification benchmarks. EDCA achieved statistically similar results in performance to FLAML and TPOT but used significantly less data to train the final solutions. Moreover, EDCA experimental results reveal that a good performance can be achieved using less data and efficient ML algorithm aspects that align with Green AutoML guidelines
Authors: Zhen Yu, Jianan Han, Yang Liu, Qingchao Chen
Abstract: Single-cell RNA sequencing (scRNA-seq) technology has profiled hundreds of millions of human cells across organs, diseases, development and perturbations to date. However, the high-dimensional sparsity, batch effect noise, category imbalance, and ever-increasing data scale of the original sequencing data pose significant challenges for multi-center knowledge transfer, data fusion, and cross-validation between scRNA-seq datasets. To address these barriers, (1) we first propose a latent codes-based scRNA-seq dataset distillation framework named scDD, which transfers and distills foundation model knowledge and original dataset information into a compact latent space and generates synthetic scRNA-seq dataset by a generator to replace the original dataset. Then, (2) we propose a single-step conditional diffusion generator named SCDG, which perform single-step gradient back-propagation to help scDD optimize distillation quality and avoid gradient decay caused by multi-step back-propagation. Meanwhile, SCDG ensures the scRNA-seq data characteristics and inter-class discriminability of the synthetic dataset through flexible conditional control and generation quality assurance. Finally, we propose a comprehensive benchmark to evaluate the performance of scRNA-seq dataset distillation in different data analysis tasks. It is validated that our proposed method can achieve 7.61% absolute and 15.70% relative improvement over previous state-of-the-art methods on average task.
Authors: Yiheng Zhu, Mingyang Li, Junlong Liu, Kun Fu, Jiansheng Wu, Qiuyi Li, Mingze Yin, Jieping Ye, Jian Wu, Zheng Wang
Abstract: Structure-based drug discovery (SBDD) is a systematic scientific process that develops new drugs by leveraging the detailed physical structure of the target protein. Recent advancements in pre-trained models for biomolecules have demonstrated remarkable success across various biochemical applications, including drug discovery and protein engineering. However, in most approaches, the pre-trained models primarily focus on the characteristics of either small molecules or proteins, without delving into their binding interactions which are essential cross-domain relationships pivotal to SBDD. To fill this gap, we propose a general-purpose foundation model named BIT (an abbreviation for Biomolecular Interaction Transformer), which is capable of encoding a range of biochemical entities, including small molecules, proteins, and protein-ligand complexes, as well as various data formats, encompassing both 2D and 3D structures. Specifically, we introduce Mixture-of-Domain-Experts (MoDE) to handle the biomolecules from diverse biochemical domains and Mixture-of-Structure-Experts (MoSE) to capture positional dependencies in the molecular structures. The proposed mixture-of-experts approach enables BIT to achieve both deep fusion and domain-specific encoding, effectively capturing fine-grained molecular interactions within protein-ligand complexes. Then, we perform cross-domain pre-training on the shared Transformer backbone via several unified self-supervised denoising tasks. Experimental results on various benchmarks demonstrate that BIT achieves exceptional performance in downstream tasks, including binding affinity prediction, structure-based virtual screening, and molecular property prediction.
Authors: Giovanni De Felice, Arianna Casanova Flores, Francesco De Santis, Silvia Santini, Johannes Schneider, Pietro Barbiero, Alberto Termine
Abstract: Concept-based models are an emerging paradigm in deep learning that constrains the inference process to operate through human-interpretable concepts, facilitating explainability and human interaction. However, these architectures, on par with popular opaque neural models, fail to account for the true causal mechanisms underlying the target phenomena represented in the data. This hampers their ability to support causal reasoning tasks, limits out-of-distribution generalization, and hinders the implementation of fairness constraints. To overcome these issues, we propose \emph{Causally reliable Concept Bottleneck Models} (C$^2$BMs), a class of concept-based architectures that enforce reasoning through a bottleneck of concepts structured according to a model of the real-world causal mechanisms. We also introduce a pipeline to automatically learn this structure from observational data and \emph{unstructured} background knowledge (e.g., scientific literature). Experimental evidence suggest that C$^2$BM are more interpretable, causally reliable, and improve responsiveness to interventions w.r.t. standard opaque and concept-based models, while maintaining their accuracy.
Authors: Antonio Guill\'en-Teruel (Departamento de Ingenier\'ia de la Informaci\'on y Las Comunicaciones, Universidad de Murcia, Murcia, 30100, Murcia, Spain), Marcos Caracena (Departamento de Ingenier\'ia de la Informaci\'on y Las Comunicaciones, Universidad de Murcia, Murcia, 30100, Murcia, Spain), Jose A. Pardo (Departamento de Ingenier\'ia de la Informaci\'on y Las Comunicaciones, Universidad de Murcia, Murcia, 30100, Murcia, Spain), Fernando de-la-G\'andara (Departamento de Ingenier\'ia de la Informaci\'on y Las Comunicaciones, Universidad de Murcia, Murcia, 30100, Murcia, Spain), Jos\'e Palma (Departamento de Ingenier\'ia de la Informaci\'on y Las Comunicaciones, Universidad de Murcia, Murcia, 30100, Murcia, Spain), Juan A. Bot\'ia (Departamento de Ingenier\'ia de la Informaci\'on y Las Comunicaciones, Universidad de Murcia, Murcia, 30100, Murcia, Spain, Department of Neurodegenerative Disease, Institute of Neurology, University College London, London, WC1N 3BG, UK)
Abstract: This research addresses the challenges of handling unbalanced datasets for binary classification tasks. In such scenarios, standard evaluation metrics are often biased by the disproportionate representation of the minority class. Conducting experiments across seven datasets, we uncovered inconsistencies in evaluation metrics when determining the model that outperforms others for each binary classification problem. This justifies the need for a metric that provides a more consistent and unbiased evaluation across unbalanced datasets, thereby supporting robust model selection. To mitigate this problem, we propose a novel metric, the Unbiased Integration Coefficients (UIC), which exhibits significantly reduced bias ($p < 10^{-4}$) towards the minority class compared to conventional metrics. The UIC is constructed by aggregating existing metrics while penalising those more prone to imbalance. In addition, we introduce the Identical Partitions for Imbalance Problems (IPIP) algorithm for imbalanced ML problems, an ensemble-based approach. Our experimental results show that IPIP outperforms other baseline imbalance-aware approaches using Random Forest and Logistic Regression models in three out of seven datasets as assessed by the UIC metric, demonstrating its effectiveness in addressing imbalanced data challenges in binary classification tasks. This new framework for dealing with imbalanced datasets is materialized in the FILM (Framework for Imbalanced Learning Machines) R Package, accessible at https://github.com/antoniogt/FILM.
Authors: Mingxue Xu, Lisa Alazraki, Danilo P. Mandic
Abstract: Pruning assumes a subnetwork exists in the original deep neural network, which can achieve comparative model performance with less computation than the original. However, it is unclear how the model performance varies with the different subnetwork extractions. In this paper, we choose the representation dimension (or embedding dimension, model dimension, the dimension of the residual stream in the relevant literature) as the entry point to this issue. We investigate the linear transformations in the LLM transformer blocks and consider a specific structured pruning approach, SliceGPT, to extract the subnetworks of different representation dimensions. We mechanistically analyse the activation flow during the model forward passes, and find the representation dimension dominates the linear transformations, model predictions, and, finally, the model performance. Explicit analytical relations are given to calculate the pruned model performance (perplexity and accuracy) without actual evaluation, and are empirically validated with Llama-3-8B-Instruct and Phi-3-mini-4k-Instruct.
Authors: Yan Li, Pengfei Zheng, Shuang Chen, Zewei Xu, Yunfei Du, Zhengang Wang
Abstract: MoE (Mixture of Experts) prevails as a neural architecture that can scale modern transformer-based LLMs (Large Language Models) to unprecedented scales. Nevertheless, large MoEs' great demands of computing power, memory capacity and memory bandwidth make scalable serving a fundamental challenge and efficient parallel inference has become a requisite to attain adequate throughput under latency constraints. DeepSpeed-MoE, one state-of-the-art MoE inference framework, adopts a 3D-parallel paradigm including EP (Expert Parallelism), TP (Tensor Parallel) and DP (Data Parallelism). However, our analysis shows DeepSpeed-MoE's inference efficiency is largely bottlenecked by EP, which is implemented with costly all-to-all collectives to route token activation. Our work aims to boost DeepSpeed-MoE by strategically reducing EP's communication overhead with a technique named Speculative MoE. Speculative MoE has two speculative parallelization schemes, speculative token shuffling and speculative expert grouping, which predict outstanding tokens' expert routing paths and pre-schedule tokens and experts across devices to losslessly trim EP's communication volume. Besides DeepSpeed-MoE, we also build Speculative MoE into a prevailing MoE inference engine SGLang. Experiments show Speculative MoE can significantly boost state-of-the-art MoE inference frameworks on fast homogeneous and slow heterogeneous interconnects.
Authors: Majed Luay, Siamak Layeghy, Seyedehfaezeh Hosseininoorbin, Mohanad Sarhan, Nour Moustafa, Marius Portmann
Abstract: This paper investigates the temporal analysis of NetFlow datasets for machine learning (ML)-based network intrusion detection systems (NIDS). Although many previous studies have highlighted the critical role of temporal features, such as inter-packet arrival time and flow length/duration, in NIDS, the currently available NetFlow datasets for NIDS lack these temporal features. This study addresses this gap by creating and making publicly available a set of NetFlow datasets that incorporate these temporal features [1]. With these temporal features, we provide a comprehensive temporal analysis of NetFlow datasets by examining the distribution of various features over time and presenting time-series representations of NetFlow features. This temporal analysis has not been previously provided in the existing literature. We also borrowed an idea from signal processing, time frequency analysis, and tested it to see how different the time frequency signal presentations (TFSPs) are for various attacks. The results indicate that many attacks have unique patterns, which could help ML models to identify them more easily.
Authors: Maxime Burchi, Radu Timofte
Abstract: The DreamerV3 algorithm recently obtained remarkable performance across diverse environment domains by learning an accurate world model based on Recurrent Neural Networks (RNNs). Following the success of model-based reinforcement learning algorithms and the rapid adoption of the Transformer architecture for its superior training efficiency and favorable scaling properties, recent works such as STORM have proposed replacing RNN-based world models with Transformer-based world models using masked self-attention. However, despite the improved training efficiency of these methods, their impact on performance remains limited compared to the Dreamer algorithm, struggling to learn competitive Transformer-based world models. In this work, we show that the next state prediction objective adopted in previous approaches is insufficient to fully exploit the representation capabilities of Transformers. We propose to extend world model predictions to longer time horizons by introducing TWISTER (Transformer-based World model wIth contraSTivE Representations), a world model using action-conditioned Contrastive Predictive Coding to learn high-level temporal feature representations and improve the agent performance. TWISTER achieves a human-normalized mean score of 162% on the Atari 100k benchmark, setting a new record among state-of-the-art methods that do not employ look-ahead search.
Authors: Marco Arazzi, Mert Cihangiroglu, Antonino Nocera
Abstract: Federated Averaging remains the most widely used aggregation strategy in federated learning due to its simplicity and scalability. However, its performance degrades significantly in non-IID data settings, where client distributions are highly imbalanced or skewed. Additionally, it relies on clients transmitting metadata, specifically the number of training samples, which introduces privacy risks and may conflict with regulatory frameworks like the European GDPR. In this paper, we propose a novel aggregation strategy that addresses these challenges by introducing class-aware gradient masking. Unlike traditional approaches, our method relies solely on gradient updates, eliminating the need for any additional client metadata, thereby enhancing privacy protection. Furthermore, our approach validates and dynamically weights client contributions based on class-specific importance, ensuring robustness against non-IID distributions, convergence prevention, and backdoor attacks. Extensive experiments on benchmark datasets demonstrate that our method not only outperforms FedAvg and other widely accepted aggregation strategies in non-IID settings but also preserves model integrity in adversarial scenarios. Our results establish the effectiveness of gradient masking as a practical and secure solution for federated learning.
Authors: Yi Shen, Jian Zhang, Jieyun Huang, Shuming Shi, Wenjing Zhang, Jiangze Yan, Ning Wang, Kai Wang, Shiguo Lian
Abstract: Recent advancements in slow-thinking reasoning models have shown exceptional performance in complex reasoning tasks. However, these models often exhibit overthinking-generating redundant reasoning steps for simple problems, leading to excessive computational resource usage. While current mitigation strategies uniformly reduce reasoning tokens, they risk degrading performance on challenging tasks that require extended reasoning. This paper introduces Difficulty-Adaptive Slow-Thinking (DAST), a novel framework that enables models to autonomously adjust the length of Chain-of-Thought(CoT) based on problem difficulty. We first propose a Token Length Budget (TLB) metric to quantify difficulty, then leveraging length-aware reward shaping and length preference optimization to implement DAST. DAST penalizes overlong responses for simple tasks while incentivizing sufficient reasoning for complex problems. Experiments on diverse datasets and model scales demonstrate that DAST effectively mitigates overthinking (reducing token usage by over 30\% on average) while preserving reasoning accuracy on complex problems.
Authors: Francisco Eiras, Eliott Zemour, Eric Lin, Vaikkunth Mugunthan
Abstract: Large Language Model (LLM) based judges form the underpinnings of key safety evaluation processes such as offline benchmarking, automated red-teaming, and online guardrailing. This widespread requirement raises the crucial question: can we trust the evaluations of these evaluators? In this paper, we highlight two critical challenges that are typically overlooked: (i) evaluations in the wild where factors like prompt sensitivity and distribution shifts can affect performance and (ii) adversarial attacks that target the judge. We highlight the importance of these through a study of commonly used safety judges, showing that small changes such as the style of the model output can lead to jumps of up to 0.24 in the false negative rate on the same dataset, whereas adversarial attacks on the model generation can fool some judges into misclassifying 100% of harmful generations as safe ones. These findings reveal gaps in commonly used meta-evaluation benchmarks and weaknesses in the robustness of current LLM judges, indicating that low attack success under certain judges could create a false sense of security.
Authors: Saif Anwar, Nathan Griffiths, Thomas Popham, Abhir Bhalerao
Abstract: Recent improvements in the expressive power of spatio-temporal models have led to performance gains in many real-world applications, such as traffic forecasting and social network modelling. However, understanding the predictions from a model is crucial to ensure reliability and trustworthiness, particularly for high-risk applications, such as healthcare and transport. Few existing methods are able to generate explanations for models trained on continuous-time dynamic graph data and, of these, the computational complexity and lack of suitable explanation objectives pose challenges. In this paper, we propose $\textbf{S}$patio-$\textbf{T}$emporal E$\textbf{X}$planation $\textbf{Search}$ (STX-Search), a novel method for generating instance-level explanations that is applicable to static and dynamic temporal graph structures. We introduce a novel search strategy and objective function, to find explanations that are highly faithful and interpretable. When compared with existing methods, STX-Search produces explanations of higher fidelity whilst optimising explanation size to maintain interpretability.
Authors: Thien Pham, Angelo Furno, Fa\"icel Chamroukhi, Latifa Oukhellou
Abstract: This paper presents an advanced Federated Learning (FL) framework for forecasting complex spatiotemporal data, improving upon recent state-of-the-art models. In the proposed approach, the original Gated Recurrent Unit (GRU) module within previous Dynamic Spatial--Temporal Graph Convolutional Recurrent Network (DSTGCRN) modeling is first replaced with a Long Short-Term Memory (LSTM) network, enabling the resulting model to more effectively capture long-term dependencies inherent to time series data. The resulting architecture significantly improves the model's capacity to handle complex temporal patterns in diverse forecasting applications. Furthermore, the proposed FL framework integrates a novel Client-Side Validation (CSV) mechanism, introducing a critical validation step at the client level before incorporating aggregated parameters from the central server into local models. This ensures that only the most effective updates are adopted, improving both the robustness and accuracy of the forecasting model across clients. The efficiency of our approach is demonstrated through extensive experiments on real-world applications, including public datasets for multimodal transport demand forecasting and private datasets for Origin-Destination (OD) matrix forecasting in urban areas. The results demonstrate substantial improvements over conventional methods, highlighting the framework's ability to capture complex spatiotemporal dependencies while preserving data privacy. This work not only provides a scalable and privacy-preserving solution for real-time, region-specific forecasting and management but also underscores the potential of leveraging distributed data sources in a FL context. We provide our algorithms as open-source on GitHub.
Authors: Louis McConnell
Abstract: In settings where both spurious and causal predictors are available, standard neural networks trained under the objective of empirical risk minimization (ERM) with no additional inductive biases tend to have a dependence on a spurious feature. As a result, it is necessary to integrate additional inductive biases in order to guide the network toward generalizable hypotheses. Often these spurious features are shared across related tasks, such as estimating disease prognoses from image scans coming from different hospitals, making the challenge of generalization more difficult. In these settings, it is important that methods are able to integrate the proper inductive biases to generalize across both nuisance-varying families as well as task families. Motivated by this setting, we present RIME (Robustly Informed Meta lEarning), a new method for meta learning under the presence of both positive and negative inductive biases (what to learn and what not to learn). We first develop a theoretical causal framework showing why existing approaches at knowledge integration can lead to worse performance on distributionally robust objectives. We then show that RIME is able to simultaneously integrate both biases, reaching state of the art performance under distributionally robust objectives in informed meta-learning settings under nuisance-varying families.
Authors: Th\'eo Gnassounou, Antoine Collas, R\'emi Flamary, Alexandre Gramfort
Abstract: Distribution shift poses a significant challenge in machine learning, particularly in biomedical applications such as EEG signals collected across different subjects, institutions, and recording devices. While existing normalization layers, Batch-Norm, LayerNorm and InstanceNorm, help address distribution shifts, they fail to capture the temporal dependencies inherent in temporal signals. In this paper, we propose PSDNorm, a layer that leverages Monge mapping and temporal context to normalize feature maps in deep learning models. Notably, the proposed method operates as a test-time domain adaptation technique, addressing distribution shifts without additional training. Evaluations on 10 sleep staging datasets using the U-Time model demonstrate that PSDNorm achieves state-of-the-art performance at test time on datasets not seen during training while being 4x more data-efficient than the best baseline. Additionally, PSDNorm provides a significant improvement in robustness, achieving markedly higher F1 scores for the 20% hardest subjects.
Authors: Manuel Santos Pereira, Lu\'is Tripa, N\'elson Lima, Francisco Caldas, Cl\'audia Soares
Abstract: First formulated by Sir Isaac Newton in his work "Philosophiae Naturalis Principia Mathematica", the concept of the Three-Body Problem was put forth as a study of the motion of the three celestial bodies within the Earth-Sun-Moon system. In a generalized definition, it seeks to predict the motion for an isolated system composed of three point masses freely interacting under Newton's law of universal attraction. This proves to be analogous to a multitude of interactions between celestial bodies, and thus, the problem finds applicability within the studies of celestial mechanics. Despite numerous attempts by renowned physicists to solve it throughout the last three centuries, no general closed-form solutions have been reached due to its inherently chaotic nature for most initial conditions. Current state-of-the-art solutions are based on two approaches, either numerical high-precision integration or machine learning-based. Notwithstanding the breakthroughs of neural networks, these present a significant limitation, which is their ignorance of any prior knowledge of the chaotic systems presented. Thus, in this work, we propose a novel method that utilizes Physics-Informed Neural Networks (PINNs). These deep neural networks are able to incorporate any prior system knowledge expressible as an Ordinary Differential Equation (ODE) into their learning processes as a regularizing agent. Our findings showcase that PINNs surpass current state-of-the-art machine learning methods with comparable prediction quality. Despite a better prediction quality, the usability of numerical integrators suffers due to their prohibitively high computational cost. These findings confirm that PINNs are both effective and time-efficient open-form solvers of the Three-Body Problem that capitalize on the extensive knowledge we hold of classical mechanics.
Authors: Yu Pan, Chaozheng Wang, Zekai Wu, Qifan Wang, Min Zhang, Zenglin Xu
Abstract: Deep neural networks have achieved remarkable accomplishments in practice. The success of these networks hinges on effective initialization methods, which are vital for ensuring stable and rapid convergence during training. Recently, initialization methods that maintain identity transition within layers have shown good efficiency in network training. These techniques (e.g., Fixup) set specific weights to zero to achieve identity control. However, settings of remaining weight (e.g., Fixup uses random values to initialize non-zero weights) will affect the inductive bias that is achieved only by a zero weight, which may be harmful to training. Addressing this concern, we introduce fully identical initialization (IDInit), a novel method that preserves identity in both the main and sub-stem layers of residual networks. IDInit employs a padded identity-like matrix to overcome rank constraints in non-square weight matrices. Furthermore, we show the convergence problem of an identity matrix can be solved by stochastic gradient descent. Additionally, we enhance the universality of IDInit by processing higher-order weights and addressing dead neuron problems. IDInit is a straightforward yet effective initialization method, with improved convergence, stability, and performance across various settings, including large-scale datasets and deep models.
Authors: Mohammad Ali Vahedifar, Qi Zhang
Abstract: Continual Learning (CL) remains a central challenge in deep learning, where models must sequentially acquire new knowledge while mitigating Catastrophic Forgetting (CF) of prior tasks. Existing approaches often struggle with efficiency and scalability, requiring extensive memory or model buffers. This work introduces ``No Forgetting Learning" (NFL), a memory-free CL framework that leverages knowledge distillation to maintain stability while preserving plasticity. Memory-free means the NFL does not rely on any memory buffer. Through extensive evaluations of three benchmark datasets, we demonstrate that NFL achieves competitive performance while utilizing approximately 14.75 times less memory than state-of-the-art methods. Furthermore, we introduce a new metric to better assess CL's plasticity-stability trade-off.
Authors: Blaine Quackenbush, Paul J. Atzberger
Abstract: We introduce methods for obtaining pretrained Geometric Neural Operators (GNPs) that can serve as basal foundation models for use in obtaining geometric features. These can be used within data processing pipelines for machine learning tasks and numerical methods. We show how our GNPs can be trained to learn robust latent representations for the differential geometry of point-clouds to provide estimates of metric, curvature, and other shape-related features. We demonstrate how our pre-trained GNPs can be used (i) to estimate the geometric properties of surfaces of arbitrary shape and topologies with robustness in the presence of noise, (ii) to approximate solutions of geometric partial differential equations (PDEs) on manifolds, and (iii) to solve equations for shape deformations such as curvature driven flows. We also release a package of the codes and weights for using our pre-trained GNPs for processing point cloud representations. This allows for incorporating our pre-trained GNPs as components for reuse within existing and new data processing pipelines. The GNPs also can be used as part of numerical solvers involving geometry or as part of methods for performing inference and other geometric tasks.
Authors: Jiang Li, Xiaoping Wang
Abstract: Protein-protein interaction (PPI) prediction is an instrumental means in elucidating the mechanisms underlying cellular operations, holding significant practical implications for the realms of pharmaceutical development and clinical treatment. Presently, the majority of research methods primarily concentrate on the analysis of amino acid sequences, while investigations predicated on protein structures remain in the nascent stages of exploration. Despite the emergence of several structure-based algorithms in recent years, these are still confronted with inherent challenges: (1) the extraction of intrinsic structural information of proteins typically necessitates the expenditure of substantial computational resources; (2) these models are overly reliant on seen protein data, struggling to effectively unearth interaction cues between unknown proteins. To further propel advancements in this domain, this paper introduces a novel PPI prediction method jointing masked reconstruction and contrastive learning, termed JmcPPI. This methodology dissects the PPI prediction task into two distinct phases: during the residue structure encoding phase, JmcPPI devises two feature reconstruction tasks and employs graph attention mechanism to capture structural information between residues; during the protein interaction inference phase, JmcPPI perturbs the original PPI graph and employs a multi-graph contrastive learning strategy to thoroughly mine extrinsic interaction information of novel proteins. Extensive experiments conducted on three widely utilized PPI datasets demonstrate that JmcPPI surpasses existing optimal baseline models across various data partition schemes. The associated code can be accessed via https://github.com/lijfrank-open/JmcPPI.
Authors: Shengzhuang Chen, Yikai Liao, Xiaoxiao Sun, Kede Ma, Ying Wei
Abstract: The advent of the foundation model era has sparked significant research interest in leveraging pre-trained representations for continual learning (CL), yielding a series of top-performing CL methods on standard evaluation benchmarks. Nonetheless, there are growing concerns regarding potential data contamination during the pre-training stage. Furthermore, standard evaluation benchmarks, which are typically static, fail to capture the complexities of real-world CL scenarios, resulting in saturated performance. To address these issues, we describe CL on dynamic benchmarks (CLDyB), a general computational framework based on Markov decision processes for evaluating CL methods reliably. CLDyB dynamically identifies inherently difficult and algorithm-dependent tasks for the given CL methods, and determines challenging task orders using Monte Carlo tree search. Leveraging CLDyB, we first conduct a joint evaluation of multiple state-of-the-art CL methods, leading to a set of commonly challenging and generalizable task sequences where existing CL methods tend to perform poorly. We then conduct separate evaluations of individual CL methods using CLDyB, discovering their respective strengths and weaknesses. The source code and generated task sequences are publicly accessible at https://github.com/szc12153/CLDyB.
Authors: Ryan Barron, Maksim E. Eren, Duc P. Truong, Cynthia Matuszek, James Wendelberger, Mary F. Dorn, Boian Alexandrov
Abstract: Missing link prediction is a method for network analysis, with applications in recommender systems, biology, social sciences, cybersecurity, information retrieval, and Artificial Intelligence (AI) reasoning in Knowledge Graphs. Missing link prediction identifies unseen but potentially existing connections in a network by analyzing the observed patterns and relationships. In proliferation detection, this supports efforts to identify and characterize attempts by state and non-state actors to acquire nuclear weapons or associated technology - a notoriously challenging but vital mission for global security. Dimensionality reduction techniques like Non-Negative Matrix Factorization (NMF) and Logistic Matrix Factorization (LMF) are effective but require selection of the matrix rank parameter, that is, of the number of hidden features, k, to avoid over/under-fitting. We introduce novel Weighted (WNMFk), Boolean (BNMFk), and Recommender (RNMFk) matrix factorization methods, along with ensemble variants incorporating logistic factorization, for link prediction. Our methods integrate automatic model determination for rank estimation by evaluating stability and accuracy using a modified bootstrap methodology and uncertainty quantification (UQ), assessing prediction reliability under random perturbations. We incorporate Otsu threshold selection and k-means clustering for Boolean matrix factorization, comparing them to coordinate descent-based Boolean thresholding. Our experiments highlight the impact of rank k selection, evaluate model performance under varying test-set sizes, and demonstrate the benefits of UQ for reliable predictions using abstention. We validate our methods on three synthetic datasets (Boolean and uniformly distributed) and benchmark them against LMF and symmetric LMF (symLMF) on five real-world protein-protein interaction networks, showcasing an improved prediction performance.
Authors: Sachit Gaudi, Gautam Sreekumar, Vishnu Boddeti
Abstract: Machine learning systems struggle with robustness, under subpopulation shifts. This problem becomes especially pronounced in scenarios where only a subset of attribute combinations is observed during training -a severe form of subpopulation shift, referred as compositional shift. To address this problem, we ask the following question: Can we improve the robustness by training on synthetic data, spanning all possible attribute combinations? We first show that training of conditional diffusion models on limited data lead to incorrect underlying distribution. Therefore, synthetic data sampled from such models will result in unfaithful samples and does not lead to improve performance of downstream machine learning systems. To address this problem, we propose CoInD to reflect the compositional nature of the world by enforcing conditional independence through minimizing Fisher's divergence between joint and marginal distributions. We demonstrate that synthetic data generated by CoInD is faithful and this translates to state-of-the-art worst-group accuracy on compositional shift tasks on CelebA.
Authors: Alireza Behtash, Marijan Fofonjka, Ethan Baird, Tyler Mauer, Hossein Moghimifam, David Stout, Joel Dennison
Abstract: We present a novel approach to selective model quantization that transcends the limitations of architecture-specific and size-dependent compression methods for Large Language Models (LLMs) using Entropy-Weighted Quantization (EWQ). By analyzing the entropy distribution across transformer blocks, EWQ determines which blocks can be safely quantized without causing significant performance degradation, independent of model architecture or size. Our method outperforms uniform quantization approaches, maintaining Massive Multitask Language Understanding (MMLU) accuracy scores within 0.5% of unquantized models while reducing memory usage by up to 18%. We demonstrate the effectiveness of EWQ across multiple architectures-from 1.6B to 70B parameters-showcasing consistent improvements in the quality-compression trade-off regardless of model scale or architectural design. A surprising finding of EWQ is its ability to reduce perplexity compared to unquantized models, suggesting the presence of beneficial regularization through selective precision reduction. This improvement holds across different model families, indicating a fundamental relationship between layer-level entropy and optimal precision requirements. Additionally, we introduce FastEWQ, a rapid method for entropy distribution analysis that eliminates the need for loading model weights. This technique leverages universal characteristics of entropy distribution that persist across various architectures and scales, enabling near-instantaneous quantization decisions while maintaining 80% classification accuracy with full entropy analysis. Our results demonstrate that effective quantization strategies can be developed independently of specific architectural choices or model sizes, opening new possibilities for efficient LLM deployment.
Authors: Udaya Ghai, Karan Singh
Abstract: Boosting provides a practical and provably effective framework for constructing accurate learning algorithms from inaccurate rules of thumb. It extends the promise of sample-efficient learning to settings where direct Empirical Risk Minimization (ERM) may not be implementable efficiently. In the realizable setting, boosting is known to offer this computational reprieve without compromising on sample efficiency. However, in the agnostic case, existing boosting algorithms fall short of achieving the optimal sample complexity. This paper highlights an unexpected and previously unexplored avenue of improvement: unlabeled samples. We design a computationally efficient agnostic boosting algorithm that matches the sample complexity of ERM, given polynomially many additional unlabeled samples. In fact, we show that the total number of samples needed, unlabeled and labeled inclusive, is never more than that for the best known agnostic boosting algorithm -- so this result is never worse -- while only a vanishing fraction of these need to be labeled for the algorithm to succeed. This is particularly fortuitous for learning-theoretic applications of agnostic boosting, which often take place in the distribution-specific setting, where unlabeled samples can be availed for free. We detail other applications of this result in reinforcement learning.
Authors: Houyi Li, Wenzheng Zheng, Jingcheng Hu, Qiufeng Wang, Hanshan Zhang, Zili Wang, Yangshijie Xu, Shuigeng Zhou, Xiangyu Zhang, Daxin Jiang
Abstract: The impressive capabilities of Large Language Models (LLMs) across diverse tasks are now well-established, yet their effective deployment necessitates careful hyperparameter optimization. Through extensive empirical studies involving grid searches across diverse configurations, we discover universal scaling laws governing these hyperparameters: optimal learning rate follows a power-law relationship with both model parameters and data sizes, while optimal batch size scales primarily with data sizes. Our analysis reveals a convex optimization landscape for hyperparameters under fixed models and data size conditions. This convexity implies an optimal hyperparameter plateau. We contribute a universal, plug-and-play optimal hyperparameter tool for the community. Its estimated values on the test set are merely 0.07\% away from the globally optimal LLM performance found via an exhaustive search. These laws demonstrate remarkable robustness across variations in model sparsity, training data distribution, and model shape. To our best known, this is the first work that unifies different model shapes and structures, such as Mixture-of-Experts models and dense transformers, as well as establishes optimal hyperparameter scaling laws across diverse data distributions. This exhaustive optimization process demands substantial computational resources, utilizing nearly one million NVIDIA H800 GPU hours to train 3,700 LLMs of varying sizes and hyperparameters from scratch and consuming approximately 100 trillion tokens in total. To facilitate reproducibility and further research, we will progressively release all loss measurements and model checkpoints through our designated repository https://step-law.github.io/
Authors: Yepeng Huang, Xiaorui Su, Varun Ullanat, Ivy Liang, Lindsay Clegg, Damilola Olabode, Nicholas Ho, Bino John, Megan Gibbs, Marinka Zitnik
Abstract: Predicting clinical outcomes from preclinical data is essential for identifying safe and effective drug combinations. Current models rely on structural or target-based features to identify high-efficacy, low-toxicity drug combinations. However, these approaches fail to incorporate the multimodal data necessary for accurate, clinically-relevant predictions. Here, we introduce MADRIGAL, a multimodal AI model that learns from structural, pathway, cell viability, and transcriptomic data to predict drug combination effects across 953 clinical outcomes and 21842 compounds, including combinations of approved drugs and novel compounds in development. MADRIGAL uses a transformer bottleneck module to unify preclinical drug data modalities while handling missing data during training and inference--a major challenge in multimodal learning. It outperforms single-modality methods and state-of-the-art models in predicting adverse drug interactions. MADRIGAL performs virtual screening of anticancer drug combinations and supports polypharmacy management for type II diabetes and metabolic dysfunction-associated steatohepatitis (MASH). It identifies transporter-mediated drug interactions. MADRIGAL predicts resmetirom, the first and only FDA-approved drug for MASH, among therapies with the most favorable safety profile. It supports personalized cancer therapy by integrating genomic profiles from cancer patients. Using primary acute myeloid leukemia samples and patient-derived xenograft models, it predicts the efficacy of personalized drug combinations. Integrating MADRIGAL with a large language model allows users to describe clinical outcomes in natural language, improving safety assessment by identifying potential adverse interactions and toxicity risks. MADRIGAL provides a multimodal approach for designing combination therapies with improved predictive accuracy and clinical relevance.
Authors: Sha Ye, Qiong Wu, Pingyi Fan, Qiang Fan
Abstract: Internet of Vehicles (IoV), as the core of intelligent transportation system, enables comprehensive interconnection between vehicles and their surroundings through multiple communication modes, which is significant for autonomous driving and intelligent traffic management. However, with the emergence of new applications, traditional communication technologies face the problems of scarce spectrum resources and high latency. Semantic communication, which focuses on extracting, transmitting, and recovering some useful semantic information from messages, can reduce redundant data transmission, improve spectrum utilization, and provide innovative solutions to communication challenges in the IoV. This paper systematically reviews state of art of semantic communications in the IoV, elaborates the technical background of IoV and semantic communications, and deeply discusses key technologies of semantic communications in IoV, including semantic information extraction, semantic communication architecture, resource allocation and management, and so on. Through specific case studies, it demonstrates that semantic communications can be effectively employed in the scenarios of traffic environment perception and understanding, intelligent driving decision support, IoV service optimization, and intelligent traffic management. Additionally, it analyzes the current challenges and future research directions. This survey reveals that semantic communications has broad application prospects in IoV, but it is necessary to solve the real existing problems by combining advanced technologies to promote its wide application in IoV and contributing to the development of intelligent transportation system.
Authors: Yiting Wei (Guangdong University of Technology, Guangzhou, China), Bingo Wing-Kuen Ling (Guangdong University of Technology, Guangzhou, China), Danni Chen (Guangdong University of Technology, Guangzhou, China), Yuheng Dai (Guangdong University of Technology, Guangzhou, China), Qing Liu (Guangdong University of Technology, Guangzhou, China)
Abstract: Recently, the wearable and non-invasive blood glucose estimation approach has been proposed. However, due to the unreliability of the acquisition device, the presence of the noise and the variations of the acquisition environments, the obtained features and the reference blood glucose values are highly unreliable. To address this issue, this paper proposes a polynomial fitting approach to smooth the obtained features or the reference blood glucose values. First, the blood glucose values are estimated based on the individual optimization approaches. Second, the absolute difference values between the estimated blood glucose values and the actual blood glucose values based on each optimization approach are computed. Third, these absolute difference values for each optimization approach are sorted in the ascending order. Fourth, for each sorted blood glucose value, the optimization method corresponding to the minimum absolute difference value is selected. Fifth, the accumulate probability of each selected optimization method is computed. If the accumulate probability of any selected optimization method at a point is greater than a threshold value, then the accumulate probabilities of these three selected optimization methods at that point are reset to zero. A range of the sorted blood glucose values are defined as that with the corresponding boundaries points being the previous reset point and this reset point. Hence, after performing the above procedures for all the sorted reference blood glucose values in the validation set, the regions of the sorted reference blood glucose values and the corresponding optimization methods in these regions are determined. The computer numerical simulation results show that our proposed method yields the mean absolute relative deviation (MARD) at 0.0930 and the percentage of the test data falling in the zone A of the Clarke error grid at 94.1176%.
Authors: Carlos Albors, Jianan Canal Li, Gonzalo Benegas, Chengzhong Ye, Yun S. Song
Abstract: Genomic language models (gLMs) have shown mostly modest success in identifying evolutionarily constrained elements in mammalian genomes. To address this issue, we introduce a novel framework for training gLMs that explicitly models nucleotide evolution on phylogenetic trees using multispecies whole-genome alignments. Our approach integrates an alignment into the loss function during training but does not require it for making predictions, thereby enhancing the model's applicability. We applied this framework to train PhyloGPN, a model that excels at predicting functionally disruptive variants from a single sequence alone and demonstrates strong transfer learning capabilities.
Authors: Shun Liao, Paolo Di Achille, Jiang Wu, Silviu Borac, Jonathan Wang, Xin Liu, Eric Teasley, Lawrence Cai, Yun Liu, Daniel McDuff, Hao-Wei Su, Brent Winslow, Anupam Pathak, Shwetak Patel, Jameson K. Rogers, Ming-Zher Poh
Abstract: Resting heart rate (RHR) is an important biomarker of cardiovascular health and mortality, but tracking it longitudinally generally requires a wearable device, limiting its availability. We present PHRM, a deep learning system for passive heart rate (HR) and RHR measurements during everyday smartphone use, using facial video-based photoplethysmography. Our system was developed using 225,773 videos from 495 participants and validated on 185,970 videos from 205 participants in laboratory and free-living conditions, representing the largest validation study of its kind. Compared to reference electrocardiogram, PHRM achieved a mean absolute percentage error (MAPE) < 10% for HR measurements across three skin tone groups of light, medium and dark pigmentation; MAPE for each skin tone group was non-inferior versus the others. Daily RHR measured by PHRM had a mean absolute error < 5 bpm compared to a wearable HR tracker, and was associated with known risk factors. These results highlight the potential of smartphones to enable passive and equitable heart health monitoring.
Authors: Ashwin Viswanathan Kannan, Madhumitha Ganesan
Abstract: Understanding cognitive flexibility and task-switching mechanisms in neural systems requires biologically plausible computational models. This tutorial presents a step-by-step approach to constructing a spiking neural network (SNN) that simulates task-switching dynamics within the cognitive control network. The model incorporates biologically realistic features, including lateral inhibition, adaptive synaptic weights through unsupervised Spike Timing-Dependent Plasticity (STDP), and precise neuronal parameterization within physiologically relevant ranges. The SNN is implemented using Leaky Integrate-and-Fire (LIF) neurons, which represent excitatory (glutamatergic) and inhibitory (GABAergic) populations. We utilize two real-world datasets as tasks, demonstrating how the network learns and dynamically switches between them. Experimental design follows cognitive psychology paradigms to analyze neural adaptation, synaptic weight modifications, and emergent behaviors such as Long-Term Potentiation (LTP), Long-Term Depression (LTD), and Task-Set Reconfiguration (TSR). Through a series of structured experiments, this tutorial illustrates how variations in task-switching intervals affect performance and multitasking efficiency. The results align with empirically observed neuronal responses, offering insights into the computational underpinnings of executive function. By following this tutorial, researchers can develop and extend biologically inspired SNN models for studying cognitive processes and neural adaptation.
Authors: Hyeonjun Kim, Kanghoon Lee, Junho Park, Jiachen Li, Jinkyoo Park
Abstract: Multi-Agent Reinforcement Learning (MARL) has shown promise in solving complex problems involving cooperation and competition among agents, such as an Unmanned Surface Vehicle (USV) swarm used in search and rescue, surveillance, and vessel protection. However, aligning system behavior with user preferences is challenging due to the difficulty of encoding expert intuition into reward functions. To address the issue, we propose a Reinforcement Learning with Human Feedback (RLHF) approach for MARL that resolves credit-assignment challenges through an Agent-Level Feedback system categorizing feedback into intra-agent, inter-agent, and intra-team types. To overcome the challenges of direct human feedback, we employ a Large Language Model (LLM) evaluator to validate our approach using feedback scenarios such as region constraints, collision avoidance, and task allocation. Our method effectively refines USV swarm policies, addressing key challenges in multi-agent systems while maintaining fairness and performance consistency.
Authors: Cristian Jimenez-Romero, Alper Yegenoglu, Christian Blum
Abstract: This work examines the integration of large language models (LLMs) into multi-agent simulations by replacing the hard-coded programs of agents with LLM-driven prompts. The proposed approach is showcased in the context of two examples of complex systems from the field of swarm intelligence: ant colony foraging and bird flocking. Central to this study is a toolchain that integrates LLMs with the NetLogo simulation platform, leveraging its Python extension to enable communication with GPT-4o via the OpenAI API. This toolchain facilitates prompt-driven behavior generation, allowing agents to respond adaptively to environmental data. For both example applications mentioned above, we employ both structured, rule-based prompts and autonomous, knowledge-driven prompts. Our work demonstrates how this toolchain enables LLMs to study self-organizing processes and induce emergent behaviors within multi-agent environments, paving the way for new approaches to exploring intelligent systems and modeling swarm intelligence inspired by natural phenomena. We provide the code, including simulation files and data at https://github.com/crjimene/swarm_gpt.
Authors: Andrew J. Larkoski
Abstract: Metrics for rigorously defining a distance between two events have been used to study the properties of the dataspace manifold of particle collider physics. The probability distribution of pairwise distances on this dataspace is unique with probability 1, and so this suggests a method to search for and identify new physics by the deviation of measurement from a null hypothesis prediction. To quantify the deviation statistically, we directly calculate the probability distribution of the number of event pairs that land in the bin a fixed distance apart. This distribution is not generically Gaussian and the ratio of the standard deviation to the mean entries in a bin scales inversely with the square-root of the number of events in the data ensemble. If the dataspace manifold exhibits some enhanced symmetry, the number of entries is Gaussian, and further fluctuations about the mean scale away like the inverse of the number of events. We define a robust measure of the non-Gaussianity of the bin-by-bin statistics of the distance distribution, and demonstrate in simulated data of jets from quantum chromodynamics sensitivity to the parton-to-hadron transition and that the manifold of events enjoys enhanced symmetries as their energy increases.
Authors: Christian K. Jespersen, Peter Melchior, David N. Spergel, Andy D. Goulding, ChangHoon Hahn, Kartheik G. Iyer
Abstract: Galaxies are often modelled as composites of separable components with distinct spectral signatures, implying that different wavelength ranges are only weakly correlated. They are not. We present a data-driven model which exploits subtle correlations between physical processes to accurately predict infrared (IR) WISE photometry from a neural summary of optical SDSS spectra. The model achieves accuracies of $\chi^2_N \approx 1$ for all photometric bands in WISE, as well as good colors. We are also able to tightly constrain typically IR-derived properties, e.g. the bolometric luminosities of AGN and dust parameters such as $\mathrm{q_{PAH}}$. We find that current SED-fitting methods are incapable of making comparable predictions, and that model misspecification often leads to correlated biases in star-formation rates and AGN luminosities. To help improve SED models, we determine what features of the optical spectrum are responsible for our improved predictions, and identify several lines (CaII, SrII, FeI, [OII] and H$\alpha$), which point to the complex chronology of star formation and chemical enrichment being incorrectly modelled.
Authors: Eliza Kosoy, Annya Dahmani, Andrew K. Lampinen, Iulia M. Comsa, Soojin Jeong, Ishita Dasgupta, Kelsey Allen
Abstract: Understanding geometry relies heavily on vision. In this work, we evaluate whether state-of-the-art vision language models (VLMs) can understand simple geometric concepts. We use a paradigm from cognitive science that isolates visual understanding of simple geometry from the many other capabilities it is often conflated with such as reasoning and world knowledge. We compare model performance with human adults from the USA, as well as with prior research on human adults without formal education from an Amazonian indigenous group. We find that VLMs consistently underperform both groups of human adults, although they succeed with some concepts more than others. We also find that VLM geometric understanding is more brittle than human understanding, and is not robust when tasks require mental rotation. This work highlights interesting differences in the origin of geometric understanding in humans and machines -- e.g. from printed materials used in formal education vs. interactions with the physical world or a combination of the two -- and a small step toward understanding these differences.
Authors: Brian Pulfer, Yury Belousov, Vitaliy Kinakh, Teddy Furon, Slava Voloshynovskiy
Abstract: The study of security in machine learning mainly focuses on downstream task-specific attacks, where the adversarial example is obtained by optimizing a loss function specific to the downstream task. At the same time, it has become standard practice for machine learning practitioners to adopt publicly available pre-trained vision foundation models, effectively sharing a common backbone architecture across a multitude of applications such as classification, segmentation, depth estimation, retrieval, question-answering and more. The study of attacks on such foundation models and their impact to multiple downstream tasks remains vastly unexplored. This work proposes a general framework that forges task-agnostic adversarial examples by maximally disrupting the feature representation obtained with foundation models. We extensively evaluate the security of the feature representations obtained by popular vision foundation models by measuring the impact of this attack on multiple downstream tasks and its transferability between models.
Authors: Shuhui Zhu, Baoxiang Wang, Sriram Ganapathi Subramanian, Pascal Poupart
Abstract: The partial alignment and conflict of autonomous agents lead to mixed-motive scenarios in many real-world applications. However, agents may fail to cooperate in practice even when cooperation yields a better outcome. One well known reason for this failure comes from non-credible commitments. To facilitate commitments among agents for better cooperation, we define Markov Commitment Games (MCGs), a variant of commitment games, where agents can voluntarily commit to their proposed future plans. Based on MCGs, we propose a learnable commitment protocol via policy gradients. We further propose incentive-compatible learning to accelerate convergence to equilibria with better social welfare. Experimental results in challenging mixed-motive tasks demonstrate faster empirical convergence and higher returns for our method compared with its counterparts. Our code is available at https://github.com/shuhui-zhu/DCL.
Authors: Muhammad Waseem, Kshitij Bhatta, Chen Li, Qing Chang
Abstract: The manufacturing industry is undergoing a transformative shift, driven by cutting-edge technologies like 5G, AI, and cloud computing. Despite these advancements, effective system control, which is crucial for optimizing production efficiency, remains a complex challenge due to the intricate, knowledge-dependent nature of manufacturing processes and the reliance on domain-specific expertise. Conventional control methods often demand heavy customization, considerable computational resources, and lack transparency in decision-making. In this work, we investigate the feasibility of using Large Language Models (LLMs), particularly GPT-4, as a straightforward, adaptable solution for controlling manufacturing systems, specifically, mobile robot scheduling. We introduce an LLM-based control framework to assign mobile robots to different machines in robot assisted serial production lines, evaluating its performance in terms of system throughput. Our proposed framework outperforms traditional scheduling approaches such as First-Come-First-Served (FCFS), Shortest Processing Time (SPT), and Longest Processing Time (LPT). While it achieves performance that is on par with state-of-the-art methods like Multi-Agent Reinforcement Learning (MARL), it offers a distinct advantage by delivering comparable throughput without the need for extensive retraining. These results suggest that the proposed LLM-based solution is well-suited for scenarios where technical expertise, computational resources, and financial investment are limited, while decision transparency and system scalability are critical concerns.
Authors: Qian Feng, David S. Martinez Lema, Jianxiang Feng, Zhaopeng Chen, Alois Knoll
Abstract: Learning dexterous manipulation from few-shot demonstrations is a significant yet challenging problem for advanced, human-like robotic systems. Dense distilled feature fields have addressed this challenge by distilling rich semantic features from 2D visual foundation models into the 3D domain. However, their reliance on neural rendering models such as Neural Radiance Fields (NeRF) or Gaussian Splatting results in high computational costs. In contrast, previous approaches based on sparse feature fields either suffer from inefficiencies due to multi-view dependencies and extensive training or lack sufficient grasp dexterity. To overcome these limitations, we propose Language-ENhanced Sparse Distilled Feature Field (LensDFF), which efficiently distills view-consistent 2D features onto 3D points using our novel language-enhanced feature fusion strategy, thereby enabling single-view few-shot generalization. Based on LensDFF, we further introduce a few-shot dexterous manipulation framework that integrates grasp primitives into the demonstrations to generate stable and highly dexterous grasps. Moreover, we present a real2sim grasp evaluation pipeline for efficient grasp assessment and hyperparameter tuning. Through extensive simulation experiments based on the real2sim pipeline and real-world experiments, our approach achieves competitive grasping performance, outperforming state-of-the-art approaches.
Authors: Ahmad Hafez, Alireza Naderi Akhormeh, Amr Hegazy, Amr Alanwar
Abstract: The deployment of Large Language Models (LLMs) in robotic systems presents unique safety challenges, particularly in unpredictable environments. Although LLMs, leveraging zero-shot learning, enhance human-robot interaction and decision-making capabilities, their inherent probabilistic nature and lack of formal guarantees raise significant concerns for safety-critical applications. Traditional model-based verification approaches often rely on precise system models, which are difficult to obtain for real-world robotic systems and may not be fully trusted due to modeling inaccuracies, unmodeled dynamics, or environmental uncertainties. To address these challenges, this paper introduces a safety assurance framework for LLM-controlled robots based on data-driven reachability analysis, a formal verification technique that ensures all possible system trajectories remain within safe operational limits. Our framework specifically investigates the problem of instructing an LLM to navigate the robot to a specified goal and assesses its ability to generate low-level control actions that successfully guide the robot safely toward that goal. By leveraging historical data to construct reachable sets of states for the robot-LLM system, our approach provides rigorous safety guarantees against unsafe behaviors without relying on explicit analytical models. We validate the framework through experimental case studies in autonomous navigation and task planning, demonstrating its effectiveness in mitigating risks associated with LLM-generated commands. This work advances the integration of formal methods into LLM-based robotics, offering a principled and practical approach to ensuring safety in next-generation autonomous systems.
Authors: Jialiang Li, Vasyl Yurchyshyn, Jason T. L. Wang, Haimin Wang, Yasser Abduallah, Khalid A. Alobaid, Chunhui Xu, Ruizhu Chen, Yan Xu
Abstract: We present a novel deep generative model, named GenMDI, to improve the temporal resolution of line-of-sight (LOS) magnetograms of solar active regions (ARs) collected by the Michelson Doppler Imager (MDI) on board the Solar and Heliospheric Observatory (SOHO). Unlike previous studies that focus primarily on spatial super-resolution of MDI magnetograms, our approach can perform temporal super-resolution, which generates and inserts synthetic data between observed MDI magnetograms, thus providing finer temporal structure and enhanced details in the LOS data. The GenMDI model employs a conditional diffusion process, which synthesizes images by considering both preceding and subsequent magnetograms, ensuring that the generated images are not only of high-quality, but also temporally coherent with the surrounding data. Experimental results show that the GenMDI model performs better than the traditional linear interpolation method, especially in ARs with dynamic evolution in magnetic fields.
Authors: Yu. Yu. Dubenskaya, A. P. Kryukov, E. O. Gres, S. P. Polyakov, E. B. Postnikov, P. A. Volchugov, A. A. Vlaskina, D. P. Zhurov
Abstract: Modern Imaging Atmospheric Cherenkov Telescopes (IACTs) generate a huge amount of data that must be classified automatically, ideally in real time. Currently, machine learning-based solutions are increasingly being used to solve classification problems. However, these classifiers require proper training data sets to work correctly. The problem with training neural networks on real IACT data is that these data need to be pre-labeled, whereas such labeling is difficult and its results are estimates. In addition, the distribution of incoming events is highly imbalanced. Firstly, there is an imbalance in the types of events, since the number of detected gamma quanta is significantly less than the number of protons. Secondly, the energy distribution of particles of the same type is also imbalanced, since high-energy particles are extremely rare. This imbalance results in poorly trained classifiers that, once trained, do not handle rare events correctly. Using only conventional Monte Carlo event simulation methods to solve this problem is possible, but extremely resource-intensive and time-consuming. To address this issue, we propose to perform data augmentation with artificially generated events of the desired type and energy using conditional generative adversarial networks (cGANs), distinguishing classes by energy values. In the paper, we describe a simple algorithm for generating balanced data sets using cGANs. Thus, the proposed neural network model produces both imbalanced data sets for physical analysis as well as balanced data sets suitable for training other neural networks.
Authors: Sreyan Ghosh, Zhifeng Kong, Sonal Kumar, S Sakshi, Jaehyeon Kim, Wei Ping, Rafael Valle, Dinesh Manocha, Bryan Catanzaro
Abstract: Understanding and reasoning over non-speech sounds and music are crucial for both humans and AI agents to interact effectively with their environments. In this paper, we introduce Audio Flamingo 2 (AF2), an Audio-Language Model (ALM) with advanced audio understanding and reasoning capabilities. AF2 leverages (i) a custom CLAP model, (ii) synthetic Audio QA data for fine-grained audio reasoning, and (iii) a multi-stage curriculum learning strategy. AF2 achieves state-of-the-art performance with only a 3B parameter small language model, surpassing large open-source and proprietary models across over 20 benchmarks. Next, for the first time, we extend audio understanding to long audio segments (30 secs to 5 mins) and propose LongAudio, a large and novel dataset for training ALMs on long audio captioning and question-answering tasks. Fine-tuning AF2 on LongAudio leads to exceptional performance on our proposed LongAudioBench, an expert annotated benchmark for evaluating ALMs on long audio understanding capabilities. We conduct extensive ablation studies to confirm the efficacy of our approach. Project Website: https://research.nvidia.com/labs/adlr/AF2/.
Authors: Wenhui Zhu, Xin Li, Xiwen Chen, Peijie Qiu, Vamsi Krishna Vasa, Xuanzhao Dong, Yanxi Chen, Natasha Lepore, Oana Dumitrascu, Yi Su, Yalin Wang
Abstract: Recently, Multimodal Large Language Models (MLLMs) have gained significant attention for their remarkable ability to process and analyze non-textual data, such as images, videos, and audio. Notably, several adaptations of general-domain MLLMs to the medical field have been explored, including LLaVA-Med. However, these medical adaptations remain insufficiently advanced in understanding and interpreting retinal images. In contrast, medical experts emphasize the importance of quantitative analyses for disease detection and interpretation. This underscores a gap between general-domain and medical-domain MLLMs: while general-domain MLLMs excel in broad applications, they lack the specialized knowledge necessary for precise diagnostic and interpretative tasks in the medical field. To address these challenges, we introduce \textit{RetinalGPT}, a multimodal conversational assistant for clinically preferred quantitative analysis of retinal images. Specifically, we achieve this by compiling a large retinal image dataset, developing a novel data pipeline, and employing customized visual instruction tuning to enhance both retinal analysis and enrich medical knowledge. In particular, RetinalGPT outperforms MLLM in the generic domain by a large margin in the diagnosis of retinal diseases in 8 benchmark retinal datasets. Beyond disease diagnosis, RetinalGPT features quantitative analyses and lesion localization, representing a pioneering step in leveraging LLMs for an interpretable and end-to-end clinical research framework. The code is available at https://github.com/Retinal-Research/RetinalGPT
Authors: Xiangxin Zhou, Yi Xiao, Haowei Lin, Xinheng He, Jiaqi Guan, Yang Wang, Qiang Liu, Feng Zhou, Liang Wang, Jianzhu Ma
Abstract: The dynamic nature of proteins, influenced by ligand interactions, is essential for comprehending protein function and progressing drug discovery. Traditional structure-based drug design (SBDD) approaches typically target binding sites with rigid structures, limiting their practical application in drug development. While molecular dynamics simulation can theoretically capture all the biologically relevant conformations, the transition rate is dictated by the intrinsic energy barrier between them, making the sampling process computationally expensive. To overcome the aforementioned challenges, we propose to use generative modeling for SBDD considering conformational changes of protein pockets. We curate a dataset of apo and multiple holo states of protein-ligand complexes, simulated by molecular dynamics, and propose a full-atom flow model (and a stochastic version), named DynamicFlow, that learns to transform apo pockets and noisy ligands into holo pockets and corresponding 3D ligand molecules. Our method uncovers promising ligand molecules and corresponding holo conformations of pockets. Additionally, the resultant holo-like states provide superior inputs for traditional SBDD approaches, playing a significant role in practical drug discovery.
Authors: Jiarong Wu, Pavel Perezhogin, David John Gagne, Brandon Reichl, Aneesh C. Subramanian, Elizabeth Thompson, Laure Zanna
Abstract: Accurately quantifying air-sea fluxes is important for understanding air-sea interactions and improving coupled weather and climate systems. This study introduces a probabilistic framework to represent the highly variable nature of air-sea fluxes, which is missing in deterministic bulk algorithms. Assuming Gaussian distributions conditioned on the input variables, we use artificial neural networks and eddy-covariance measurement data to estimate the mean and variance by minimizing negative log-likelihood loss. The trained neural networks provide alternative mean flux estimates to existing bulk algorithms, and quantify the uncertainty around the mean estimates. Stochastic parameterization of air-sea turbulent fluxes can be constructed by sampling from the predicted distributions. Tests in a single-column forced upper-ocean model suggest that changes in flux algorithms influence sea surface temperature and mixed layer depth seasonally. The ensemble spread in stochastic runs is most pronounced during spring restratification.
Authors: Amin Karimi, Charalambos Poullis
Abstract: Few-shot semantic segmentation (FSS) aims to enable models to segment novel/unseen object classes using only a limited number of labeled examples. However, current FSS methods frequently struggle with generalization due to incomplete and biased feature representations, especially when support images do not capture the full appearance variability of the target class. To improve the FSS pipeline, we propose a novel framework that utilizes large language models (LLMs) to adapt general class semantic information to the query image. Furthermore, the framework employs dense pixel-wise matching to identify similarities between query and support images, resulting in enhanced FSS performance. Inspired by reasoning-based segmentation frameworks, our method, named DSV-LFS, introduces an additional token into the LLM vocabulary, allowing a multimodal LLM to generate a "semantic prompt" from class descriptions. In parallel, a dense matching module identifies visual similarities between the query and support images, generating a "visual prompt". These prompts are then jointly employed to guide the prompt-based decoder for accurate segmentation of the query image. Comprehensive experiments on the benchmark datasets Pascal-$5^{i}$ and COCO-$20^{i}$ demonstrate that our framework achieves state-of-the-art performance-by a significant margin-demonstrating superior generalization to novel classes and robustness across diverse scenarios. The source code is available at \href{https://github.com/aminpdik/DSV-LFS}{https://github.com/aminpdik/DSV-LFS}
URLs: https://github.com/aminpdik/DSV-LFS, https://github.com/aminpdik/DSV-LFS
Authors: Philip Charles, Deep Ray, Yue Yu, Joost Prins, Hugo Melchers, Michael R. A. Abdelmalik, Jeffrey Cochran, Assad A. Oberai, Thomas J. R. Hughes, Mats G. Larson
Abstract: The optimal Petrov-Galerkin formulation to solve partial differential equations (PDEs) recovers the best approximation in a specified finite-dimensional (trial) space with respect to a suitable norm. However, the recovery of this optimal solution is contingent on being able to construct the optimal weighting functions associated with the trial basis. While explicit constructions are available for simple one- and two-dimensional problems, such constructions for a general multidimensional problem remain elusive. In the present work, we revisit the optimal Petrov-Galerkin formulation through the lens of deep learning. We propose an operator network framework called Petrov-Galerkin Variationally Mimetic Operator Network (PG-VarMiON), which emulates the optimal Petrov-Galerkin weak form of the underlying PDE. The PG-VarMiON is trained in a supervised manner using a labeled dataset comprising the PDE data and the corresponding PDE solution, with the training loss depending on the choice of the optimal norm. The special architecture of the PG-VarMiON allows it to implicitly learn the optimal weighting functions, thus endowing the proposed operator network with the ability to generalize well beyond the training set. We derive approximation error estimates for PG-VarMiON, highlighting the contributions of various error sources, particularly the error in learning the true weighting functions. Several numerical results are presented for the advection-diffusion equation to demonstrate the efficacy of the proposed method. By embedding the Petrov-Galerkin structure into the network architecture, PG-VarMiON exhibits greater robustness and improved generalization compared to other popular deep operator frameworks, particularly when the training data is limited.
Authors: Tanumoy Banerjee, Kevin Ji, Weiyi Xia, Gaoyuan Ouyang, Tyler Del Rose, Ihor Z. Hlova, Benjamin Ueland, Duane D. Johnson, Cai-Zhuan Wang, Ganesh Balasubramanian, Prashant Singh
Abstract: The transition to a low-carbon economy demands efficient and sustainable energy-storage solutions, with hydrogen emerging as a promising clean-energy carrier and with metal hydrides recognized for their hydrogen-storage capacity. Here, we leverage machine learning (ML) to predict hydrogen-to-metal (H/M) ratios and solution energy by incorporating thermodynamic parameters and local lattice distortion (LLD) as key features. Our best-performing ML model provides improvements to H/M ratios and solution energies over a broad class of ternary alloys (easily extendable to multi-principal-element alloys), such as Ti-Nb-X (X = Mo, Cr, Hf, Ta, V, Zr) and Co-Ni-X (X = Al, Mg, V). Ti-Nb-Mo alloys reveal compositional effects in H-storage behavior, in particular Ti, Nb, and V enhance H-storage capacity, while Mo reduces H/M and hydrogen weight percent by 40-50%. We attributed to slow hydrogen kinetics in molybdenum rich alloys, which is validated by our pressure-composition isotherm (PCT) experiments on pure Ti and Ti5Mo95 alloys. Density functional theory (DFT) and molecular simulations also confirm that Ti and Nb promote H diffusion, whereas Mo hinders it, highlighting the interplay between electronic structure, lattice distortions, and hydrogen uptake. Notably, our Gradient Boosting Regression model identifies LLD as a critical factor in H/M predictions. To aid material selection, we present two periodic tables illustrating elemental effects on (a) H2 wt% and (b) solution energy, derived from ML, and provide a reference for identifying alloying elements that enhance hydrogen solubility and storage.
Authors: Xinyue Cui, Johnny Tian-Zheng Wei, Swabha Swayamdipta, Robin Jia
Abstract: Data watermarking in language models injects traceable signals, such as specific token sequences or stylistic patterns, into copyrighted text, allowing copyright holders to track and verify training data ownership. Previous data watermarking techniques primarily focus on effective memorization after pretraining, while overlooking challenges that arise in other stages of the LLM pipeline, such as the risk of watermark filtering during data preprocessing, or potential forgetting through post-training, or verification difficulties due to API-only access. We propose a novel data watermarking approach that injects coherent and plausible yet fictitious knowledge into training data using generated passages describing a fictitious entity and its associated attributes. Our watermarks are designed to be memorized by the LLM through seamlessly integrating in its training data, making them harder to detect lexically during preprocessing.We demonstrate that our watermarks can be effectively memorized by LLMs, and that increasing our watermarks' density, length, and diversity of attributes strengthens their memorization. We further show that our watermarks remain robust throughout LLM development, maintaining their effectiveness after continual pretraining and supervised finetuning. Finally, we show that our data watermarks can be evaluated even under API-only access via question answering.
Authors: Muheng Li, Ruqi Zhang
Abstract: Recently, gradient-based discrete sampling has emerged as a highly efficient, general-purpose solver for various combinatorial optimization (CO) problems, achieving performance comparable to or surpassing the popular data-driven approaches. However, we identify a critical issue in these methods, which we term ''wandering in contours''. This behavior refers to sampling new different solutions that share very similar objective values for a long time, leading to computational inefficiency and suboptimal exploration of potential solutions. In this paper, we introduce a novel reheating mechanism inspired by the concept of critical temperature and specific heat in physics, aimed at overcoming this limitation. Empirically, our method demonstrates superiority over existing sampling-based and data-driven algorithms across a diverse array of CO problems.
Authors: Karthik Elamvazhuthi
Abstract: In this note, we revisit the problem of flow approximation properties of neural ordinary differential equations (NODEs). The approximation properties have been considered as a flow controllability problem in recent literature. The neural ODE is considered {\it narrow} when the parameters have dimension equal to the input of the neural network, and hence have limited width. We derive the relation of narrow NODEs in approximating flows of shallow but wide NODEs. Due to existing results on approximation properties of shallow neural networks, this facilitates understanding which kind of flows of dynamical systems can be approximated using narrow neural ODEs. While approximation properties of narrow NODEs have been established in literature, the proofs often involve extensive constructions or require invoking deep controllability theorems from control theory. In this paper, we provide a simpler proof technique that involves only ideas from ODEs and Gr{\"o}nwall's lemma. Moreover, we provide an estimate on the number of switches needed for the time dependent weights of the narrow NODE to mimic the behavior of a NODE with a single layer wide neural network as the velocity field.
Authors: Miao Li, Michael Klamkin, Mathieu Tanneau, Reza Zandehshahvar, Pascal Van Hentenryck
Abstract: This paper studies a Conformal Prediction (CP) methodology for building prediction intervals in a regression setting, given only deterministic lower and upper bounds on the target variable. It proposes a new CP mechanism (CPUL) that goes beyond post-processing by adopting a model selection approach over multiple nested interval construction methods. Paradoxically, many well-established CP methods, including CPUL, may fail to provide adequate coverage in regions where the bounds are tight. To remedy this limitation, the paper proposes an optimal thresholding mechanism, OMLT, that adjusts CPUL intervals in tight regions with undercoverage. The combined CPUL-OMLT is validated on large-scale learning tasks where the goal is to bound the optimal value of a parametric optimization problem. The experimental results demonstrate substantial improvements over baseline methods across various datasets.
Authors: Arash Mozhdehi, Yunli Wang, Sun Sun, Xin Wang
Abstract: Deep reinforcement learning (DRL)-based frameworks, featuring Transformer-style policy networks, have demonstrated their efficacy across various vehicle routing problem (VRP) variants. However, the application of these methods to the multi-trip time-dependent vehicle routing problem (MTTDVRP) with maximum working hours constraints -- a pivotal element of urban logistics -- remains largely unexplored. This paper introduces a DRL-based method called the Simultaneous Encoder and Dual Decoder Attention Model (SED2AM), tailored for the MTTDVRP with maximum working hours constraints. The proposed method introduces a temporal locality inductive bias to the encoding module of the policy networks, enabling it to effectively account for the time-dependency in travel distance or time. The decoding module of SED2AM includes a vehicle selection decoder that selects a vehicle from the fleet, effectively associating trips with vehicles for functional multi-trip routing. Additionally, this decoding module is equipped with a trip construction decoder leveraged for constructing trips for the vehicles. This policy model is equipped with two classes of state representations, fleet state and routing state, providing the information needed for effective route construction in the presence of maximum working hours constraints. Experimental results using real-world datasets from two major Canadian cities not only show that SED2AM outperforms the current state-of-the-art DRL-based and metaheuristic-based baselines but also demonstrate its generalizability to solve larger-scale problems.
Authors: Sheikh Moonwara Anjum Monisha, Ratun Rahman
Abstract: One of the primary challenges in medical diagnostics is the accurate and efficient use of magnetic resonance imaging (MRI) for the detection of brain tumors. But the current machine learning (ML) approaches have two major limitations, data privacy and high latency. To solve the problem, in this work we propose a federated learning architecture for a better accurate brain tumor detection incorporating the YOLOv11 algorithm. In contrast to earlier methods of centralized learning, our federated learning approach protects the underlying medical data while supporting cooperative deep learning model training across multiple institutions. To allow the YOLOv11 model to locate and identify tumor areas, we adjust it to handle MRI data. To ensure robustness and generalizability, the model is trained and tested on a wide range of MRI data collected from several anonymous medical facilities. The results indicate that our method significantly maintains higher accuracy than conventional approaches.
Authors: Ziqiao Wang, Cheng Long, Yongyi Mao
Abstract: Federated Learning (FL) is a widely adopted privacy-preserving distributed learning framework, yet its generalization performance remains less explored compared to centralized learning. In FL, the generalization error consists of two components: the out-of-sample gap, which measures the gap between the empirical and true risk for participating clients, and the participation gap, which quantifies the risk difference between participating and non-participating clients. In this work, we apply an information-theoretic analysis via the conditional mutual information (CMI) framework to study FL's two-level generalization. Beyond the traditional supersample-based CMI framework, we introduce a superclient construction to accommodate the two-level generalization setting in FL. We derive multiple CMI-based bounds, including hypothesis-based CMI bounds, illustrating how privacy constraints in FL can imply generalization guarantees. Furthermore, we propose fast-rate evaluated CMI bounds that recover the best-known convergence rate for two-level FL generalization in the small empirical risk regime. For specific FL model aggregation strategies and structured loss functions, we refine our bounds to achieve improved convergence rates with respect to the number of participating clients. Empirical evaluations confirm that our evaluated CMI bounds are non-vacuous and accurately capture the generalization behavior of FL algorithms.
Authors: Erik Jones, Arjun Patrawala, Jacob Steinhardt
Abstract: Humans often rely on subjective natural language to direct language models (LLMs); for example, users might instruct the LLM to write an enthusiastic blogpost, while developers might train models to be helpful and harmless using LLM-based edits. The LLM's operational semantics of such subjective phrases -- how it adjusts its behavior when each phrase is included in the prompt -- thus dictates how aligned it is with human intent. In this work, we uncover instances of misalignment between LLMs' actual operational semantics and what humans expect. Our method, TED (thesaurus error detector), first constructs a thesaurus that captures whether two phrases have similar operational semantics according to the LLM. It then elicits failures by unearthing disagreements between this thesaurus and a human-constructed reference. TED routinely produces surprising instances of misalignment; for example, Mistral 7B Instruct produces more harassing outputs when it edits text to be witty, and Llama 3 8B Instruct produces dishonest articles when instructed to make the articles enthusiastic. Our results demonstrate that humans can uncover unexpected LLM behavior by scrutinizing relationships between abstract concepts, without supervising outputs directly.
Authors: Alan Luo, Kaiwen Yuan
Abstract: Vision Transformers (ViTs) have demonstrated exceptional performance in various vision tasks. However, they tend to underperform on smaller datasets due to their inherent lack of inductive biases. Current approaches address this limitation implicitly-often by pairing ViTs with pretext tasks or by distilling knowledge from convolutional neural networks (CNNs) to strengthen the prior. In contrast, Self-Organizing Maps (SOMs), a widely adopted self-supervised framework, are inherently structured to preserve topology and spatial organization, making them a promising candidate to directly address the limitations of ViTs in limited or small training datasets. Despite this potential, equipping SOMs with modern deep learning architectures remains largely unexplored. In this study, we conduct a novel exploration on how Vision Transformers (ViTs) and Self-Organizing Maps (SOMs) can empower each other, aiming to bridge this critical research gap. Our findings demonstrate that these architectures can synergistically enhance each other, leading to significantly improved performance in both unsupervised and supervised tasks. Code will be publicly available.
Authors: Tao Zhong, Christine Allen-Blanchette
Abstract: We propose GAGrasp, a novel framework for dexterous grasp generation that leverages geometric algebra representations to enforce equivariance to SE(3) transformations. By encoding the SE(3) symmetry constraint directly into the architecture, our method improves data and parameter efficiency while enabling robust grasp generation across diverse object poses. Additionally, we incorporate a differentiable physics-informed refinement layer, which ensures that generated grasps are physically plausible and stable. Extensive experiments demonstrate the model's superior performance in generalization, stability, and adaptability compared to existing methods. Additional details at https://gagrasp.github.io/
Authors: Jie Liu, Tiexin Qin, Hui Liu, Yilei Shi, Lichao Mou, Xiao Xiang Zhu, Shiqi Wang, Haoliang Li
Abstract: In this work, we address the challenge of adaptive pediatric Left Ventricular Ejection Fraction (LVEF) assessment. While Test-time Training (TTT) approaches show promise for this task, they suffer from two significant limitations. Existing TTT works are primarily designed for classification tasks rather than continuous value regression, and they lack mechanisms to handle the quasi-periodic nature of cardiac signals. To tackle these issues, we propose a novel \textbf{Q}uasi-\textbf{P}eriodic \textbf{A}daptive \textbf{R}egression with \textbf{T}est-time Training (Q-PART) framework. In the training stage, the proposed Quasi-Period Network decomposes the echocardiogram into periodic and aperiodic components within latent space by combining parameterized helix trajectories with Neural Controlled Differential Equations. During inference, our framework further employs a variance minimization strategy across image augmentations that simulate common quality issues in echocardiogram acquisition, along with differential adaptation rates for periodic and aperiodic components. Theoretical analysis is provided to demonstrate that our variance minimization objective effectively bounds the regression error under mild conditions. Furthermore, extensive experiments across three pediatric age groups demonstrate that Q-PART not only significantly outperforms existing approaches in pediatric LVEF prediction, but also exhibits strong clinical screening capability with high mAUROC scores (up to 0.9747) and maintains gender-fair performance across all metrics, validating its robustness and practical utility in pediatric echocardiography analysis.
Authors: Jie Xu, Na Zhao, Gang Niu, Masashi Sugiyama, Xiaofeng Zhu
Abstract: Recently, multi-view learning (MVL) has garnered significant attention due to its ability to fuse discriminative information from multiple views. However, real-world multi-view datasets are often heterogeneous and imperfect, which usually makes MVL methods designed for specific combinations of views lack application potential and limits their effectiveness. To address this issue, we propose a novel robust MVL method (namely RML) with simultaneous representation fusion and alignment. Specifically, we introduce a simple yet effective multi-view transformer fusion network where we transform heterogeneous multi-view data into homogeneous word embeddings, and then integrate multiple views by the sample-level attention mechanism to obtain a fused representation. Furthermore, we propose a simulated perturbation based multi-view contrastive learning framework that dynamically generates the noise and unusable perturbations for simulating imperfect data conditions. The simulated noisy and unusable data obtain two distinct fused representations, and we utilize contrastive learning to align them for learning discriminative and robust representations. Our RML is self-supervised and can also be applied for downstream tasks as a regularization. In experiments, we employ it in unsupervised multi-view clustering, noise-label classification, and as a plug-and-play module for cross-modal hashing retrieval. Extensive comparison experiments and ablation studies validate the effectiveness of RML.
Authors: Binghui Wu, Dinil Mon Divakaran, Mohan Gurusamy
Abstract: As modern networks grow increasingly complex--driven by diverse devices, encrypted protocols, and evolving threats--network traffic analysis has become critically important. Existing machine learning models often rely only on a single representation of packets or flows, limiting their ability to capture the contextual relationships essential for robust analysis. Furthermore, task-specific architectures for supervised, semi-supervised, and unsupervised learning lead to inefficiencies in adapting to varying data formats and security tasks. To address these gaps, we propose UniNet, a unified framework that introduces a novel multi-granular traffic representation (T-Matrix), integrating session, flow, and packet-level features to provide comprehensive contextual information. Combined with T-Attent, a lightweight attention-based model, UniNet efficiently learns latent embeddings for diverse security tasks. Extensive evaluations across four key network security and privacy problems--anomaly detection, attack classification, IoT device identification, and encrypted website fingerprinting--demonstrate UniNet's significant performance gain over state-of-the-art methods, achieving higher accuracy, lower false positive rates, and improved scalability. By addressing the limitations of single-level models and unifying traffic analysis paradigms, UniNet sets a new benchmark for modern network security.
Authors: Hejie Cui, Alyssa Unell, Bowen Chen, Jason Alan Fries, Emily Alsentzer, Sanmi Koyejo, Nigam Shah
Abstract: Large language models (LLMs) have emerged as promising tools for assisting in medical tasks, yet processing Electronic Health Records (EHRs) presents unique challenges due to their longitudinal nature. While LLMs' capabilities to perform medical tasks continue to improve, their ability to reason over temporal dependencies across multiple patient visits and time frames remains unexplored. We introduce TIMER (Temporal Instruction Modeling and Evaluation for Longitudinal Clinical Records), a framework that incorporate instruction-response pairs grounding to different parts of a patient's record as a critical dimension in both instruction evaluation and tuning for longitudinal clinical records. We develop TIMER-Bench, the first time-aware benchmark that evaluates temporal reasoning capabilities over longitudinal EHRs, as well as TIMER-Instruct, an instruction-tuning methodology for LLMs to learn reasoning over time. We demonstrate that models fine-tuned with TIMER-Instruct improve performance by 7.3% on human-generated benchmarks and 9.2% on TIMER-Bench, indicating that temporal instruction-tuning improves model performance for reasoning over EHR.
Authors: Evgeniy Eremin
Abstract: In modern world the importance of cybersecurity of various systems is increasing from year to year. The number of information security events generated by information security tools grows up with the development of the IT infrastructure. At the same time, the cyber threat landscape does not remain constant, and monitoring should take into account both already known attack indicators and those for which there are no signature rules in information security products of various classes yet. Detecting anomalies in large cybersecurity data streams is a complex task that, if properly addressed, can allow for timely response to atypical and previously unknown cyber threats. The possibilities of using of offline algorithms may be limited for a number of reasons related to the time of training and the frequency of retraining. Using stream learning algorithms for solving this task is capable of providing near-real-time data processing. This article examines the results of ten algorithms from three Python stream machine-learning libraries on BETH dataset with cybersecurity events, which contains information about the creation, cloning, and destruction of operating system processes collected using extended eBPF. ROC-AUC metric and total processing time of processing with these algorithms are presented. Several combinations of features and the order of events are considered. In conclusion, some mentions are given about the most promising algorithms and possible directions for further research are outlined.
Authors: Angelos Assos, Yuval Dagan, Nived Rajaraman
Abstract: Online learning algorithms are widely used in strategic multi-agent settings, including repeated auctions, contract design, and pricing competitions, where agents adapt their strategies over time. A key question in such environments is how an optimizing agent can best respond to a learning agent to improve its own long-term outcomes. While prior work has developed efficient algorithms for the optimizer in special cases - such as structured auction settings or contract design - no general efficient algorithm is known. In this paper, we establish a strong computational hardness result: unless $\mathsf{P} = \mathsf{NP}$, no polynomial-time optimizer can compute a near-optimal strategy against a learner using a standard no-regret algorithm, specifically Multiplicative Weights Update (MWU). Our result proves an $\Omega(T)$ hardness bound, significantly strengthening previous work that only showed an additive $\Theta(1)$ impossibility result. Furthermore, while the prior hardness result focused on learners using fictitious play - an algorithm that is not no-regret - we prove intractability for a widely used no-regret learning algorithm. This establishes a fundamental computational barrier to finding optimal strategies in general game-theoretic settings.
Authors: Zhanhong Jiang, Md Zahid Hasan, Aditya Balu, Joshua R. Waite, Genyi Huang, Soumik Sarkar
Abstract: Stochastic optimization methods have actively been playing a critical role in modern machine learning algorithms to deliver decent performance. While numerous works have proposed and developed diverse approaches, first-order and second-order methods are in entirely different situations. The former is significantly pivotal and dominating in emerging deep learning but only leads convergence to a stationary point. However, second-order methods are less popular due to their computational intensity in large-dimensional problems. This paper presents a novel method that leverages both the first-order and second-order methods in a unified algorithmic framework, termed FUSE, from which a practical version (PV) is derived accordingly. FUSE-PV stands as a simple yet efficient optimization method involving a switch-over between first and second orders. Additionally, we develop different criteria that determine when to switch. FUSE-PV has provably shown a smaller computational complexity than SGD and Adam. To validate our proposed scheme, we present an ablation study on several simple test functions and show a comparison with baselines for benchmark datasets.
Authors: Bin Wu, Wuxuan Shi, Jinqiao Wang, Mang Ye
Abstract: Pre-trained Vision-Language Models (VLMs) require Continual Learning (CL) to efficiently update their knowledge and adapt to various downstream tasks without retraining from scratch. However, for VLMs, in addition to the loss of knowledge previously learned from downstream tasks, pre-training knowledge is also corrupted during continual fine-tuning. This issue is exacerbated by the unavailability of original pre-training data, leaving VLM's generalization ability degrading. In this paper, we propose GIFT, a novel continual fine-tuning approach that utilizes synthetic data to overcome catastrophic forgetting in VLMs. Taking advantage of recent advances in text-to-image synthesis, we employ a pre-trained diffusion model to recreate both pre-training and learned downstream task data. In this way, the VLM can revisit previous knowledge through distillation on matching diffusion-generated images and corresponding text prompts. Leveraging the broad distribution and high alignment between synthetic image-text pairs in VLM's feature space, we propose a contrastive distillation loss along with an image-text alignment constraint. To further combat in-distribution overfitting and enhance distillation performance with limited amount of generated data, we incorporate adaptive weight consolidation, utilizing Fisher information from these synthetic image-text pairs and achieving a better stability-plasticity balance. Extensive experiments demonstrate that our method consistently outperforms previous state-of-the-art approaches across various settings.
Authors: Julian Aron Prenner, Romain Robbes
Abstract: Modern Large Language Models (LLMs) have shown astounding capabilities of code understanding and synthesis. In order to assess such capabilities, several benchmarks have been devised (e.g., HumanEval). However, most benchmarks focus on code synthesis from natural language instructions. Hence, such benchmarks do not test for other forms of code understanding. Moreover, there have been concerns about contamination and leakage. That is, benchmark problems (or closely related problems) may appear in training set, strongly biasing benchmark results. In this work we investigate whether large language models can correctly predict runtime program behavior. To this end, we introduce ThrowBench, a benchmark consisting of over 2,400 short user-written programs written in four different programming languages. The majority of these programs throw an exception during runtime (due to a bug). LLMs are asked to predict whether a presented program throws an exception and, if so, which one. Evaluating our benchmark on six state-of-the-art code LLMs we see modest performance ranging from 19 to 38% (F1 score). Benchmarking a wider set of code capabilities could improve the assessment of code LLMs and help identify weak points in current models. Moreover, as ground-truth answers have been determined through program execution, leakage is not a concern. We release ThrowBench as well as all of our results together with this work.
Authors: Biao Ouyang, Yingying Zhang, Hanyin Cheng, Yang Shu, Chenjuan Guo, Bin Yang, Qingsong Wen, Lunting Fan, Christian S. Jensen
Abstract: With the continued migration of storage to cloud database systems,the impact of slow queries in such systems on services and user experience is increasing. Root-cause diagnosis plays an indispensable role in facilitating slow-query detection and revision. This paper proposes a method capable of both identifying possible root cause types for slow queries and ranking these according to their potential for accelerating slow queries. This enables prioritizing root causes with the highest impact, in turn improving slow-query revision effectiveness. To enable more accurate and detailed diagnoses, we propose the multimodal Ranking for the Root Causes of slow queries (RCRank) framework, which formulates root cause analysis as a multimodal machine learning problem and leverages multimodal information from query statements, execution plans, execution logs, and key performance indicators. To obtain expressive embeddings from its heterogeneous multimodal input, RCRank integrates self-supervised pre-training that enhances cross-modal alignment and task relevance. Next, the framework integrates root-cause-adaptive cross Transformers that enable adaptive fusion of multimodal features with varying characteristics. Finally, the framework offers a unified model that features an impact-aware training objective for identifying and ranking root causes. We report on experiments on real and synthetic datasets, finding that RCRank is capable of consistently outperforming the state-of-the-art methods at root cause identification and ranking according to a range of metrics.
Authors: Niccol\`o Turcato, Matteo Iovino, Aris Synodinos, Alberto Dalla Libera, Ruggero Carli, Pietro Falco
Abstract: Recent advancements in Large Language Models (LLMs) and Visual Language Models (VLMs) have significantly impacted robotics, enabling high-level semantic motion planning applications. Reinforcement Learning (RL), a complementary paradigm, enables agents to autonomously optimize complex behaviors through interaction and reward signals. However, designing effective reward functions for RL remains challenging, especially in real-world tasks where sparse rewards are insufficient and dense rewards require elaborate design. In this work, we propose Autonomous Reinforcement learning for Complex HumanInformed Environments (ARCHIE), an unsupervised pipeline leveraging GPT-4, a pre-trained LLM, to generate reward functions directly from natural language task descriptions. The rewards are used to train RL agents in simulated environments, where we formalize the reward generation process to enhance feasibility. Additionally, GPT-4 automates the coding of task success criteria, creating a fully automated, one-shot procedure for translating human-readable text into deployable robot skills. Our approach is validated through extensive simulated experiments on single-arm and bi-manual manipulation tasks using an ABB YuMi collaborative robot, highlighting its practicality and effectiveness. Tasks are demonstrated on the real robot setup.
Authors: Ziqing Yang, Yixin Wu, Yun Shen, Wei Dai, Michael Backes, Yang Zhang
Abstract: The tremendous commercial potential of large language models (LLMs) has heightened concerns about their unauthorized use. Third parties can customize LLMs through fine-tuning and offer only black-box API access, effectively concealing unauthorized usage and complicating external auditing processes. This practice not only exacerbates unfair competition, but also violates licensing agreements. In response, identifying the origin of black-box LLMs is an intrinsic solution to this issue. In this paper, we first reveal the limitations of state-of-the-art passive and proactive identification methods with experiments on 30 LLMs and two real-world black-box APIs. Then, we propose the proactive technique, PlugAE, which optimizes adversarial token embeddings in a continuous space and proactively plugs them into the LLM for tracing and identification. The experiments show that PlugAE can achieve substantial improvement in identifying fine-tuned derivatives. We further advocate for legal frameworks and regulations to better address the challenges posed by the unauthorized use of LLMs.
Authors: Ivan Oleksiyuk, Svyatoslav Voloshynovskiy, Tobias Golling
Abstract: We introduce a new model for conditional and continuous data morphing called TRansport Adversarial Network for Smooth InTerpolation (TRANSIT). We apply it to create a background data template for weakly-supervised searches at the LHC. The method smoothly transforms sideband events to match signal region mass distributions. We demonstrate the performance of TRANSIT using the LHC Olympics R\&D dataset. The model captures non-linear mass correlations of features and produces a template that offers a competitive anomaly sensitivity compared to state-of-the-art transport-based template generators. Moreover, the computational training time required for TRANSIT is an order of magnitude lower than that of competing deep learning methods. This makes it ideal for analyses that iterate over many signal regions and signal models. Unlike generative models, which must learn a full probability density distribution, i.e., the correlations between all the variables, the proposed transport model only has to learn a smooth conditional shift of the distribution. This allows for a simpler, more efficient residual architecture, enabling mass uncorrelated features to pass the network unchanged while the mass correlated features are adjusted accordingly. Furthermore, we show that the latent space of the model provides a set of mass decorrelated features useful for anomaly detection without background sculpting.
Authors: Homer Durand, Gherardo Varando, Gustau Camps-Valls
Abstract: We propose a novel approach for learning causal response representations. Our method aims to extract directions in which a multidimensional outcome is most directly caused by a treatment variable. By bridging conditional independence testing with causal representation learning, we formulate an optimisation problem that maximises the evidence against conditional independence between the treatment and outcome, given a conditioning set. This formulation employs flexible regression models tailored to specific applications, creating a versatile framework. The problem is addressed through a generalised eigenvalue decomposition. We show that, under mild assumptions, the distribution of the largest eigenvalue can be bounded by a known $F$-distribution, enabling testable conditional independence. We also provide theoretical guarantees for the optimality of the learned representation in terms of signal-to-noise ratio and Fisher information maximisation. Finally, we demonstrate the empirical effectiveness of our approach in simulation and real-world experiments. Our results underscore the utility of this framework in uncovering direct causal effects within complex, multivariate settings.
Authors: Zhilin Wang, Jiaqi Zeng, Olivier Delalleau, Daniel Egert, Ellie Evans, Hoo-Chang Shin, Felipe Soares, Yi Dong, Oleksii Kuchaiev
Abstract: Inference-Time Scaling has been critical to the success of recent models such as OpenAI o1 and DeepSeek R1. However, many techniques used to train models for inference-time scaling require tasks to have answers that can be verified, limiting their application to domains such as math, coding and logical reasoning. We take inspiration from how humans make first attempts, ask for detailed feedback from others and make improvements based on such feedback across a wide spectrum of open-ended endeavors. To this end, we collect data for and train dedicated Feedback and Edit Models that are capable of performing inference-time scaling for open-ended general-domain tasks. In our setup, one model generates an initial response, which are given feedback by a second model, that are then used by a third model to edit the response. We show that performance on Arena Hard, a benchmark strongly predictive of Chatbot Arena Elo can be boosted by scaling the number of initial response drafts, effective feedback and edited responses. When scaled optimally, our setup based on 70B models from the Llama 3 family can reach SoTA performance on Arena Hard at 92.7 as of 5 Mar 2025, surpassing OpenAI o1-preview-2024-09-12 with 90.4 and DeepSeek R1 with 92.3.
Authors: Yiyong Luo, Brooks Paige, Jim Griffin
Abstract: Recent economic events, including the global financial crisis and COVID-19 pandemic, have exposed limitations in linear Factor Augmented Vector Autoregressive (FAVAR) models for forecasting and structural analysis. Nonlinear dimension techniques, particularly autoencoders, have emerged as promising alternatives in a FAVAR framework, but challenges remain in identifiability, interpretability, and integration with traditional nonlinear time series methods. We address these challenges through two contributions. First, we introduce a Grouped Sparse autoencoder that employs the Spike-and-Slab Lasso prior, with parameters under this prior being shared across variables of the same economic category, thereby achieving semi-identifiability and enhancing model interpretability. Second, we incorporate time-varying parameters into the VAR component to better capture evolving economic dynamics. Our empirical application to the US economy demonstrates that the Grouped Sparse autoencoder produces more interpretable factors through its parsimonious structure; and its combination with time-varying parameter VAR shows superior performance in both point and density forecasting. Impulse response analysis reveals that monetary policy shocks during recessions generate more moderate responses with higher uncertainty compared to expansionary periods.
Authors: Yu-Seung Roh, Joo-Young Kim, Jin-Duk Park, Won-Yong Shin
Abstract: Multimodal recommender systems improve the performance of canonical recommender systems with no item features by utilizing diverse content types such as text, images, and videos, while alleviating inherent sparsity of user-item interactions and accelerating user engagement. However, current neural network-based models often incur significant computational overhead due to the complex training process required to learn and integrate information from multiple modalities. To overcome this limitation, we propose MultiModal-Graph Filtering (MM-GF), a training-free method based on the notion of graph filtering (GF) for efficient and accurate multimodal recommendations. Specifically, MM-GF first constructs multiple similarity graphs through nontrivial multimodal feature refinement such as robust scaling and vector shifting by addressing the heterogeneous characteristics across modalities. Then, MM-GF optimally fuses multimodal information using linear low-pass filters across different modalities. Extensive experiments on real-world benchmark datasets demonstrate that MM-GF not only improves recommendation accuracy by up to 13.35% compared to the best competitor but also dramatically reduces computational costs by achieving the runtime of less than 10 seconds.
Authors: Xiaoqi Wang, Hongyang Du, Yuehong Gao, Dong In Kim
Abstract: Recent advancements in large language models (LLMs) have led to their widespread adoption and large-scale deployment across various domains. However, their environmental impact, particularly during inference, has become a growing concern due to their substantial energy consumption and carbon footprint. Existing research has focused on inference computation alone, overlooking the analysis and optimization of carbon footprint in network-aided LLM service systems. To address this gap, we propose AOLO, a framework for analysis and optimization for low-carbon oriented wireless LLM services. AOLO introduces a comprehensive carbon footprint model that quantifies greenhouse gas emissions across the entire LLM service chain, including computational inference and wireless communication. Furthermore, we formulate an optimization problem aimed at minimizing the overall carbon footprint, which is solved through joint optimization of inference outputs and transmit power under quality-of-experience and system performance constraints. To achieve this joint optimization, we leverage the energy efficiency of spiking neural networks (SNNs) by adopting SNN as the actor network and propose a low-carbon-oriented optimization algorithm, i.e., SNN-based deep reinforcement learning (SDRL). Comprehensive simulations demonstrate that SDRL algorithm significantly reduces overall carbon footprint, achieving an 18.77% reduction compared to the benchmark soft actor-critic, highlighting its potential for enabling more sustainable LLM inference services.
Authors: Siavash Ameli, Chris van der Heide, Liam Hodgkinson, Fred Roosta, Michael W. Mahoney
Abstract: Calculating or accurately estimating log-determinants of large positive semi-definite matrices is of fundamental importance in many machine learning tasks. While its cubic computational complexity can already be prohibitive, in modern applications, even storing the matrices themselves can pose a memory bottleneck. To address this, we derive a novel hierarchical algorithm based on block-wise computation of the LDL decomposition for large-scale log-determinant calculation in memory-constrained settings. In extreme cases where matrices are highly ill-conditioned, accurately computing the full matrix itself may be infeasible. This is particularly relevant when considering kernel matrices at scale, including the empirical Neural Tangent Kernel (NTK) of neural networks trained on large datasets. Under the assumption of neural scaling laws in the test error, we show that the ratio of pseudo-determinants satisfies a power-law relationship, allowing us to derive corresponding scaling laws. This enables accurate estimation of NTK log-determinants from a tiny fraction of the full dataset; in our experiments, this results in a $\sim$100,000$\times$ speedup with improved accuracy over competing approximations. Using these techniques, we successfully estimate log-determinants for dense matrices of extreme sizes, which were previously deemed intractable and inaccessible due to their enormous scale and computational demands.
Authors: Natalia Cherezova, Artur Jutman, Maksim Jenihhin
Abstract: The emergence of Deep Neural Networks (DNNs) in mission- and safety-critical applications brings their reliability to the front. High performance demands of DNNs require the use of specialized hardware accelerators. Systolic array architecture is widely used in DNN accelerators due to its parallelism and regular structure. This work presents a run-time reconfigurable systolic array architecture with three execution modes and four implementation options. All four implementations are evaluated in terms of resource utilization, throughput, and fault tolerance improvement. The proposed architecture is used for reliability enhancement of DNN inference on systolic array through heterogeneous mapping of different network layers to different execution modes. The approach is supported by a novel reliability assessment method based on fault propagation analysis. It is used for the exploration of the appropriate execution mode-layer mapping for DNN inference. The proposed architecture efficiently protects registers and MAC units of systolic array PEs from transient and permanent faults. The reconfigurability feature enables a speedup of up to $3\times$, depending on layer vulnerability. Furthermore, it requires $6\times$ less resources compared to static redundancy and $2.5\times$ less resources compared to the previously proposed solution for transient faults.
Authors: Wei Liu, Xin Liu, Michael K. Ng, Zaikun Zhang
Abstract: Semi-supervised clustering is a basic problem in various applications. Most existing methods require knowledge of the ideal cluster number, which is often difficult to obtain in practice. Besides, satisfying the must-link constraints is another major challenge for these methods. In this work, we view the semi-supervised clustering task as a partitioning problem on a graph associated with the given dataset, where the similarity matrix includes a scaling parameter to reflect the must-link constraints. Utilizing a relaxation technique, we formulate the graph partitioning problem into a continuous optimization model that does not require the exact cluster number, but only an overestimate of it. We then propose a block coordinate descent algorithm to efficiently solve this model, and establish its convergence result. Based on the obtained solution, we can construct the clusters that theoretically meet the must-link constraints under mild assumptions. Furthermore, we verify the effectiveness and efficiency of our proposed method through comprehensive numerical experiments.
Authors: Runhan Chen, Meijin Lin, Jianshu Chen, Liangjie Lin, Jiazheng Wang, Xiaoqing Li, Jianhua Wang, Xu Huang, Ling Qian, Shaoxing Liu, Yuan Long, Di Guo, Xiaobo Qu, Haiwei Han
Abstract: Given the need to elucidate the mechanisms underlying illnesses and their treatment, as well as the lack of harmonization of acquisition and post-processing protocols among different magnetic resonance system vendors, this work is to determine if metabolite concentrations obtained from different sessions, machine models and even different vendors of 3 T scanners can be highly reproducible and be pooled for diagnostic analysis, which is very valuable for the research of rare diseases. Participants underwent magnetic resonance imaging (MRI) scanning once on two separate days within one week (one session per day, each session including two proton magnetic resonance spectroscopy (1H-MRS) scans with no more than a 5-minute interval between scans (no off-bed activity)) on each machine. were analyzed for reliability of within- and between- sessions using the coefficient of variation (CV) and intraclass correlation coefficient (ICC), and for reproducibility of across the machines using correlation coefficient. As for within- and between- session, all CV values for a group of all the first or second scans of a session, or for a session were almost below 20%, and most of the ICCs for metabolites range from moderate (0.4-0.59) to excellent (0.75-1), indicating high data reliability. When it comes to the reproducibility across the three scanners, all Pearson correlation coefficients across the three machines approached 1 with most around 0.9, and majority demonstrated statistical significance (P<0.01). Additionally, the intra-vendor reproducibility was greater than the inter-vendor ones.
Authors: Xiangyu Miao, Jun Sun, Hang Lai, Xinpeng Di, Jiahang Cao, Yong Yu, Weinan Zhang
Abstract: With the rapid development of embodied intelligence, locomotion control of quadruped robots on complex terrains has become a research hotspot. Unlike traditional locomotion control approaches focusing solely on velocity tracking, we pursue to balance the agility and robustness of quadruped robots on diverse and complex terrains. To this end, we propose an end-to-end deep reinforcement learning framework for posture-aware locomotion named PALo, which manages to handle simultaneous linear and angular velocity tracking and real-time adjustments of body height, pitch, and roll angles. In PALo, the locomotion control problem is formulated as a partially observable Markov decision process, and an asymmetric actor-critic architecture is adopted to overcome the sim-to-real challenge. Further, by incorporating customized training curricula, PALo achieves agile posture-aware locomotion control in simulated environments and successfully transfers to real-world settings without fine-tuning, allowing real-time control of the quadruped robot's locomotion and body posture across challenging terrains. Through in-depth experimental analysis, we identify the key components of PALo that contribute to its performance, further validating the effectiveness of the proposed method. The results of this study provide new possibilities for the low-level locomotion control of quadruped robots in higher dimensional command spaces and lay the foundation for future research on upper-level modules for embodied intelligence.
Authors: Meijin Lin, Lin Guo, Dicheng Chen, Jianshu Chen, Zhangren Tu, Xu Huang, Jianhua Wang, Ji Qi, Yuan Long, Zhiguo Huang, Di Guo, Xiaobo Qu, Haiwei Han
Abstract: Objctives: This work aimed to statistically compare the metabolite quantification of human brain magnetic resonance spectroscopy (MRS) between the deep learning method QNet and the classical method LCModel through an easy-to-use intelligent cloud computing platform CloudBrain-MRS. Materials and Methods: In this retrospective study, two 3 T MRI scanners Philips Ingenia and Achieva collected 61 and 46 in vivo 1H magnetic resonance (MR) spectra of healthy participants, respectively, from the brain region of pregenual anterior cingulate cortex from September to October 2021. The analyses of Bland-Altman, Pearson correlation and reasonability were performed to assess the degree of agreement, linear correlation and reasonability between the two quantification methods. Results: Fifteen healthy volunteers (12 females and 3 males, age range: 21-35 years, mean age/standard deviation = 27.4/3.9 years) were recruited. The analyses of Bland-Altman, Pearson correlation and reasonability showed high to good consistency and very strong to moderate correlation between the two methods for quantification of total N-acetylaspartate (tNAA), total choline (tCho), and inositol (Ins) (relative half interval of limits of agreement = 3.04%, 9.3%, and 18.5%, respectively; Pearson correlation coefficient r = 0.775, 0.927, and 0.469, respectively). In addition, quantification results of QNet are more likely to be closer to the previous reported average values than those of LCModel. Conclusion: There were high or good degrees of consistency between the quantification results of QNet and LCModel for tNAA, tCho, and Ins, and QNet generally has more reasonable quantification than LCModel.
Authors: Matthieu Carreau, Roi Naveiro, William N. Caballero
Abstract: Research in adversarial machine learning (AML) has shown that statistical models are vulnerable to maliciously altered data. However, despite advances in Bayesian machine learning models, most AML research remains concentrated on classical techniques. Therefore, we focus on extending the white-box model poisoning paradigm to attack generic Bayesian inference, highlighting its vulnerability in adversarial contexts. A suite of attacks are developed that allow an attacker to steer the Bayesian posterior toward a target distribution through the strategic deletion and replication of true observations, even when only sampling access to the posterior is available. Analytic properties of these algorithms are proven and their performance is empirically examined in both synthetic and real-world scenarios. With relatively little effort, the attacker is able to substantively alter the Bayesian's beliefs and, by accepting more risk, they can mold these beliefs to their will. By carefully constructing the adversarial posterior, surgical poisoning is achieved such that only targeted inferences are corrupted and others are minimally disturbed.
Authors: Dimitri von R\"utte, Janis Fluri, Yuhui Ding, Antonio Orvieto, Bernhard Sch\"olkopf, Thomas Hofmann
Abstract: While state-of-the-art language models achieve impressive results through next-token prediction, they have inherent limitations such as the inability to revise already generated tokens. This has prompted exploration of alternative approaches such as discrete diffusion. However, masked diffusion, which has emerged as a popular choice due to its simplicity and effectiveness, reintroduces this inability to revise words. To overcome this, we generalize masked diffusion and derive the theoretical backbone of a family of general interpolating discrete diffusion (GIDD) processes offering greater flexibility in the design of the noising processes. Leveraging a novel diffusion ELBO, we achieve compute-matched state-of-the-art performance in diffusion language modeling. Exploiting GIDD's flexibility, we explore a hybrid approach combining masking and uniform noise, leading to improved sample quality and unlocking the ability for the model to correct its own mistakes, an area where autoregressive models notoriously have struggled. Our code and models are open-source: https://github.com/dvruette/gidd/
Authors: Tianyu Cui, Song-Jun Xu, Artem Moskalev, Shuwei Li, Tommaso Mansi, Mangal Prakash, Rui Liao
Abstract: Inferring Gene Regulatory Networks (GRNs) from gene expression data is crucial for understanding biological processes. While supervised models are reported to achieve high performance for this task, they rely on costly ground truth (GT) labels and risk learning gene-specific biases, such as class imbalances of GT interactions, rather than true regulatory mechanisms. To address these issues, we introduce InfoSEM, an unsupervised generative model that leverages textual gene embeddings as informative priors, improving GRN inference without GT labels. InfoSEM can also integrate GT labels as an additional prior when available, avoiding biases and further enhancing performance. Additionally, we propose a biologically motivated benchmarking framework that better reflects real-world applications such as biomarker discovery and reveals learned biases of existing supervised methods. InfoSEM outperforms existing models by 38.5% across four datasets using textual embeddings prior and further boosts performance by 11.1% when integrating labeled data as priors.
Authors: Joohwi Lee, Kaito Miyamoto
Abstract: In the rapidly advancing field of materials informatics, nonlinear machine learning models have demonstrated exceptional predictive capabilities for material properties. However, their black-box nature limits interpretability, and they may incorporate features that do not contribute to, or even deteriorate, model performance. This study employs explainable ML (XML) techniques, including permutation feature importance and the SHapley Additive exPlanation, applied to a pristine support vector regression model designed to predict band gaps at the GW level using 18 input features. Guided by XML-derived individual feature importance, a simple framework is proposed to construct reduced-feature predictive models. Model evaluations indicate that an XML-guided compact model, consisting of the top five features, achieves comparable accuracy to the pristine model on in-domain datasets while demonstrating superior generalization with lower prediction errors on out-of-domain data. Additionally, the study underscores the necessity for eliminating strongly correlated features to prevent misinterpretation and overestimation of feature importance before applying XML. This study highlights XML's effectiveness in developing simplified yet highly accurate machine learning models by clarifying feature roles.
Authors: Adrian Chang, Kai Wang, Yuanbo Li, Manolis Savva, Angel X. Chang, Daniel Ritchie
Abstract: Data driven and autoregressive indoor scene synthesis systems generate indoor scenes automatically by suggesting and then placing objects one at a time. Empirical observations show that current systems tend to produce incomplete next object location distributions. We introduce a system which addresses this problem. We design a Domain Specific Language (DSL) that specifies functional constraints. Programs from our language take as input a partial scene and object to place. Upon execution they predict possible object placements. We design a generative model which writes these programs automatically. Available 3D scene datasets do not contain programs to train on, so we build upon previous work in unsupervised program induction to introduce a new program bootstrapping algorithm. In order to quantify our empirical observations we introduce a new evaluation procedure which captures how well a system models per-object location distributions. We ask human annotators to label all the possible places an object can go in a scene and show that our system produces per-object location distributions more consistent with human annotators. Our system also generates indoor scenes of comparable quality to previous systems and while previous systems degrade in performance when training data is sparse, our system does not degrade to the same degree.
Authors: Alexander M. Tanaka, Aras T. Asaad, Richard Cooper, Vidit Nanda
Abstract: We introduce a new ligand-based virtual screening (LBVS) framework that uses piecewise linear (PL) Morse theory to predict ligand binding potential. We model ligands as simplicial complexes via a pruned Delaunay triangulation, and catalogue the critical points across multiple directional height functions. This produces a rich feature vector, consisting of crucial topological features -- peaks, troughs, and saddles -- that characterise ligand surfaces relevant to binding interactions. Unlike contemporary LBVS methods that rely on computationally-intensive deep neural networks, we require only a lightweight classifier. The Morse theoretic approach achieves state-of-the-art performance on standard datasets while offering an interpretable feature vector and scalable method for ligand prioritization in early-stage drug discovery.
Authors: Sumit Vashishtha, Odalric-Ambrym Maillard
Abstract: We introduce Dirichlet Process Posterior Sampling (DPPS), a Bayesian non-parametric algorithm for multi-arm bandits based on Dirichlet Process (DP) priors. Like Thompson-sampling, DPPS is a probability-matching algorithm, i.e., it plays an arm based on its posterior-probability of being optimal. Instead of assuming a parametric class for the reward generating distribution of each arm, and then putting a prior on the parameters, in DPPS the reward generating distribution is directly modeled using DP priors. DPPS provides a principled approach to incorporate prior belief about the bandit environment, and in the noninformative limit of the DP posteriors (i.e. Bayesian Bootstrap), we recover Non Parametric Thompson Sampling (NPTS), a popular non-parametric bandit algorithm, as a special case of DPPS. We employ stick-breaking representation of the DP priors, and show excellent empirical performance of DPPS in challenging synthetic and real world bandit environments. Finally, using an information-theoretic analysis, we show non-asymptotic optimality of DPPS in the Bayesian regret setup.
Authors: Jacqueline R. M. A. Maasch, Alihan H\"uy\"uk, Xinnuo Xu, Aditya V. Nori, Javier Gonzalez
Abstract: Causal reasoning and compositional reasoning are two core aspirations in generative AI. Measuring the extent of these behaviors requires principled evaluation methods. We explore a unified perspective that considers both behaviors simultaneously, termed compositional causal reasoning (CCR): the ability to infer how causal measures compose and, equivalently, how causal quantities propagate through graphs. We instantiate a framework for the systematic evaluation of CCR for the average treatment effect and the probability of necessity and sufficiency. As proof of concept, we demonstrate the design of CCR tasks for language models in the LLama, Phi, and GPT families. On a math word problem, our framework revealed a range of taxonomically distinct error patterns. Additionally, CCR errors increased with the complexity of causal paths for all models except o1.
Authors: Rafael I. Cabral Muchacho, Florian T. Pokorny
Abstract: Geodesic distances play a fundamental role in robotics, as they efficiently encode global geometric information of the domain. Recent methods use neural networks to approximate geodesic distances by solving the Eikonal equation through physics-informed approaches. While effective, these approaches often suffer from unstable convergence during training in complex environments. We propose a framework to learn geodesic distances in irregular domains by using the Soner boundary condition, and systematically evaluate the impact of data losses on training stability and solution accuracy. Our experiments demonstrate that incorporating data losses significantly improves convergence robustness, reducing training instabilities and sensitivity to initialization. These findings suggest that hybrid data-physics approaches can effectively enhance the reliability of learning-based geodesic distance solvers with sparse data.
Authors: Zhijian Zhuo, Yutao Zeng, Ya Wang, Sijun Zhang, Jian Yang, Xiaoqing Li, Xun Zhou, Jinwen Ma
Abstract: Transformers have become the de facto architecture for a wide range of machine learning tasks, particularly in large language models (LLMs). Despite their remarkable performance, challenges remain in training deep transformer networks, especially regarding the location of layer normalization. While Pre-Norm structures facilitate easier training due to their more prominent identity path, they often yield suboptimal performance compared to Post-Norm. In this paper, we propose $\textbf{HybridNorm}$, a straightforward yet effective hybrid normalization strategy that integrates the advantages of both Pre-Norm and Post-Norm approaches. Specifically, HybridNorm employs QKV normalization within the attention mechanism and Post-Norm in the feed-forward network (FFN) of each transformer block. This design not only stabilizes training but also enhances performance, particularly in the context of LLMs. Comprehensive experiments in both dense and sparse architectures show that HybridNorm consistently outperforms both Pre-Norm and Post-Norm approaches, achieving state-of-the-art results across various benchmarks. These findings highlight the potential of HybridNorm as a more stable and effective technique for improving the training and performance of deep transformer models. %Code will be made publicly available. Code is available at https://github.com/BryceZhuo/HybridNorm.
Authors: Aoxiong Yin, Kai Shen, Yichong Leng, Xu Tan, Xinyu Zhou, Juncheng Li, Siliang Tang
Abstract: Recent advancements in text-to-video (T2V) generation have been driven by two competing paradigms: autoregressive language models and diffusion models. However, each paradigm has intrinsic limitations: language models struggle with visual quality and error accumulation, while diffusion models lack semantic understanding and causal modeling. In this work, we propose LanDiff, a hybrid framework that synergizes the strengths of both paradigms through coarse-to-fine generation. Our architecture introduces three key innovations: (1) a semantic tokenizer that compresses 3D visual features into compact 1D discrete representations through efficient semantic compression, achieving a $\sim$14,000$\times$ compression ratio; (2) a language model that generates semantic tokens with high-level semantic relationships; (3) a streaming diffusion model that refines coarse semantics into high-fidelity videos. Experiments show that LanDiff, a 5B model, achieves a score of 85.43 on the VBench T2V benchmark, surpassing the state-of-the-art open-source models Hunyuan Video (13B) and other commercial models such as Sora, Keling, and Hailuo. Furthermore, our model also achieves state-of-the-art performance in long video generation, surpassing other open-source models in this field. Our demo can be viewed at https://landiff.github.io/.
Authors: Yijie Xu, Aiwei Liu, Xuming Hu, Lijie Wen, Hui Xiong
Abstract: As open-source large language models (LLMs) like Llama3 become more capable, it is crucial to develop watermarking techniques to detect their potential misuse. Existing watermarking methods either add watermarks during LLM inference, which is unsuitable for open-source LLMs, or primarily target classification LLMs rather than recent generative LLMs. Adapting these watermarks to open-source LLMs for misuse detection remains an open challenge. This work defines two misuse scenarios for open-source LLMs: intellectual property (IP) violation and LLM Usage Violation. Then, we explore the application of inference-time watermark distillation and backdoor watermarking in these contexts. We propose comprehensive evaluation methods to assess the impact of various real-world further fine-tuning scenarios on watermarks and the effect of these watermarks on LLM performance. Our experiments reveal that backdoor watermarking could effectively detect IP Violation, while inference-time watermark distillation is applicable in both scenarios but less robust to further fine-tuning and has a more significant impact on LLM performance compared to backdoor watermarking. Exploring more advanced watermarking methods for open-source LLMs to detect their misuse should be an important future direction.
Authors: Aishik Konwer, Zhijian Yang, Erhan Bas, Cao Xiao, Prateek Prasanna, Parminder Bhatia, Taha Kass-Hout
Abstract: Foundational models such as the Segment Anything Model (SAM) are gaining traction in medical imaging segmentation, supporting multiple downstream tasks. However, such models are supervised in nature, still relying on large annotated datasets or prompts supplied by experts. Conventional techniques such as active learning to alleviate such limitations are limited in scope and still necessitate continuous human involvement and complex domain knowledge for label refinement or establishing reward ground truth. To address these challenges, we propose an enhanced Segment Anything Model (SAM) framework that utilizes annotation-efficient prompts generated in a fully unsupervised fashion, while still capturing essential semantic, location, and shape information through contrastive language-image pretraining and visual question answering. We adopt the direct preference optimization technique to design an optimal policy that enables the model to generate high-fidelity segmentations with simple ratings or rankings provided by a virtual annotator simulating the human annotation process. State-of-the-art performance of our framework in tasks such as lung segmentation, breast tumor segmentation, and organ segmentation across various modalities, including X-ray, ultrasound, and abdominal CT, justifies its effectiveness in low-annotation data scenarios.
Authors: Yuqi Hu, Longguang Wang, Xian Liu, Ling-Hao Chen, Yuwei Guo, Yukai Shi, Ce Liu, Anyi Rao, Zeyu Wang, Hui Xiong
Abstract: Understanding and replicating the real world is a critical challenge in Artificial General Intelligence (AGI) research. To achieve this, many existing approaches, such as world models, aim to capture the fundamental principles governing the physical world, enabling more accurate simulations and meaningful interactions. However, current methods often treat different modalities, including 2D (images), videos, 3D, and 4D representations, as independent domains, overlooking their interdependencies. Additionally, these methods typically focus on isolated dimensions of reality without systematically integrating their connections. In this survey, we present a unified survey for multimodal generative models that investigate the progression of data dimensionality in real-world simulation. Specifically, this survey starts from 2D generation (appearance), then moves to video (appearance+dynamics) and 3D generation (appearance+geometry), and finally culminates in 4D generation that integrate all dimensions. To the best of our knowledge, this is the first attempt to systematically unify the study of 2D, video, 3D and 4D generation within a single framework. To guide future research, we provide a comprehensive review of datasets, evaluation metrics and future directions, and fostering insights for newcomers. This survey serves as a bridge to advance the study of multimodal generative models and real-world simulation within a unified framework.
Authors: Dou Hu, Lingwei Wei, Wei Zhou, Songlin Hu
Abstract: This paper proposes a new principled multi-task representation learning framework (InfoMTL) to extract noise-invariant sufficient representations for all tasks. It ensures sufficiency of shared representations for all tasks and mitigates the negative effect of redundant features, which can enhance language understanding of pre-trained language models (PLMs) under the multi-task paradigm. Firstly, a shared information maximization principle is proposed to learn more sufficient shared representations for all target tasks. It can avoid the insufficiency issue arising from representation compression in the multi-task paradigm. Secondly, a task-specific information minimization principle is designed to mitigate the negative effect of potential redundant features in the input for each task. It can compress task-irrelevant redundant information and preserve necessary information relevant to the target for multi-task prediction. Experiments on six classification benchmarks show that our method outperforms 12 comparative multi-task methods under the same multi-task settings, especially in data-constrained and noisy scenarios. Extensive experiments demonstrate that the learned representations are more sufficient, data-efficient, and robust.
Authors: Nathaniel Haynam, Adam Khoja, Dhruv Kumar, Vivek Myers, Erdem B{\i}y{\i}k
Abstract: When reward functions are hand-designed, deep reinforcement learning algorithms often suffer from reward misspecification, causing them to learn suboptimal policies in terms of the intended task objectives. In the single-agent case, inverse reinforcement learning (IRL) techniques attempt to address this issue by inferring the reward function from expert demonstrations. However, in multi-agent problems, misalignment between the learned and true objectives is exacerbated due to increased environment non-stationarity and variance that scales with multiple agents. As such, in multi-agent general-sum games, multi-agent IRL algorithms have difficulty balancing cooperative and competitive objectives. To address these issues, we propose Multi-Agent Marginal Q-Learning from Demonstrations (MAMQL), a novel sample-efficient framework for multi-agent IRL. For each agent, MAMQL learns a critic marginalized over the other agents' policies, allowing for a well-motivated use of Boltzmann policies in the multi-agent context. We identify a connection between optimal marginalized critics and single-agent soft-Q IRL, allowing us to apply a direct, simple optimization criterion from the single-agent domain. Across our experiments on three different simulated domains, MAMQL significantly outperforms previous multi-agent methods in average reward, sample efficiency, and reward recovery by often more than 2-5x. We make our code available at https://sites.google.com/view/mamql .
Authors: Dingling Yao, Filip Tronarp, Nathanael Bosch
Abstract: Filtering-based probabilistic numerical solvers for ordinary differential equations (ODEs), also known as ODE filters, have been established as efficient methods for quantifying numerical uncertainty in the solution of ODEs. In practical applications, however, the underlying dynamical system often contains uncertain parameters, requiring the propagation of this model uncertainty to the ODE solution. In this paper, we demonstrate that ODE filters, despite their probabilistic nature, do not automatically solve this uncertainty propagation problem. To address this limitation, we present a novel approach that combines ODE filters with numerical quadrature to properly marginalize over uncertain parameters, while accounting for both parameter uncertainty and numerical solver uncertainty. Experiments across multiple dynamical systems demonstrate that the resulting uncertainty estimates closely match reference solutions. Notably, we show how the numerical uncertainty from the ODE solver can help prevent overconfidence in the propagated uncertainty estimates, especially when using larger step sizes. Our results illustrate that probabilistic numerical methods can effectively quantify both numerical and parametric uncertainty in dynamical systems.
Authors: Ike Griss Salas, Megan R. Ebers, Jake Stevens-Haas, J. Nathan Kutz
Abstract: Monitoring the atmospheric dispersion of pollutants is increasingly critical for environmental impact assessments. High-fidelity computational models are often employed to simulate plume dynamics, guiding decision-making and prioritizing resource deployment. However, such models can be prohibitively expensive to simulate, as they require resolving turbulent flows at fine spatial and temporal resolutions. Moreover, there are at least two distinct dynamical regimes of interest in the plume: (i) the initial ejection of the plume where turbulent mixing is generated by the shear-driven Kelvin-Helmholtz instability, and (ii) the ensuing turbulent diffusion and advection which is often modeled by the Gaussian plume model. We address the challenge of modeling the initial plume generation. Specifically, we propose a data-driven framework that identifies a reduced-order analytical model for plume dynamics -- directly from video data. We extract a time series of plume center and edge points from video snapshots and evaluate different regressions based to their extrapolation performance to generate a time series of coefficients that characterize the plume's overall direction and spread. We regress to a sinusoidal model inspired by the Kelvin-Helmholtz instability for the edge points in order to identify the plume's dispersion and vorticity. Overall, this reduced-order modeling framework provides a data-driven and lightweight approach to capture the dominant features of the initial nonlinear point-source plume dynamics, agnostic to plume type and starting only from video. The resulting model is a pre-cursor to standard models such as the Gaussian plume model and has the potential to enable rapid assessment and evaluation of critical environmental hazards, such as methane leaks, chemical spills, and pollutant dispersal from smokestacks.
Authors: Pranjal Aggarwal, Sean Welleck
Abstract: Reasoning language models have shown an uncanny ability to improve performance at test-time by ``thinking longer''-that is, by generating longer chain-of-thought sequences and hence using more compute. However, the length of their chain-of-thought reasoning is not controllable, making it impossible to allocate test-time compute to achieve a desired level of performance. We introduce Length Controlled Policy Optimization (LCPO), a simple reinforcement learning method that optimizes for accuracy and adherence to user-specified length constraints. We use LCPO to train L1, a reasoning language model that produces outputs satisfying a length constraint given in its prompt. L1's length control allows for smoothly trading off computational cost and accuracy on a wide range of tasks, and outperforms the state-of-the-art S1 method for length control. Furthermore, we uncover an unexpected short chain-of-thought capability in models trained with LCPO. For instance, our 1.5B L1 model surpasses GPT-4o at equal reasoning lengths. Overall, LCPO enables precise control over reasoning length, allowing for fine-grained allocation of test-time compute and accuracy. We release code and models at https://www.cmu-l3.github.io/l1
Authors: Daniel Yiming Cao, August Y. Chen, Karthik Sridharan, Benjamin Tang
Abstract: In this paper, we study the problem of non-convex optimization on functions that are not necessarily smooth using first order methods. Smoothness (functions whose gradient and/or Hessian are Lipschitz) is not satisfied by many machine learning problems in both theory and practice, motivating a recent line of work studying the convergence of first order methods to first order stationary points under appropriate generalizations of smoothness. We develop a novel framework to study convergence of first order methods to first and \textit{second} order stationary points under generalized smoothness, under more general smoothness assumptions than the literature. Using our framework, we show appropriate variants of GD and SGD (e.g. with appropriate perturbations) can converge not just to first order but also \textit{second order stationary points} in runtime polylogarithmic in the dimension. To our knowledge, our work contains the first such result, as well as the first 'non-textbook' rate for non-convex optimization under generalized smoothness. We demonstrate that several canonical non-convex optimization problems fall under our setting and framework.
Authors: Anuj Diwan, Zhisheng Zheng, David Harwath, Eunsol Choi
Abstract: We introduce Paralinguistic Speech Captions (ParaSpeechCaps), a large-scale dataset that annotates speech utterances with rich style captions. While rich abstract tags (e.g. guttural, nasal, pained) have been explored in small-scale human-annotated datasets, existing large-scale datasets only cover basic tags (e.g. low-pitched, slow, loud). We combine off-the-shelf text and speech embedders, classifiers and an audio language model to automatically scale rich tag annotations for the first time. ParaSpeechCaps covers a total of 59 style tags, including both speaker-level intrinsic tags and utterance-level situational tags. It consists of 342 hours of human-labelled data (PSC-Base) and 2427 hours of automatically annotated data (PSC-Scaled). We finetune Parler-TTS, an open-source style-prompted TTS model, on ParaSpeechCaps, and achieve improved style consistency (+7.9% Consistency MOS) and speech quality (+15.5% Naturalness MOS) over the best performing baseline that combines existing rich style tag datasets. We ablate several of our dataset design choices to lay the foundation for future work in this space. Our dataset, models and code are released at https://github.com/ajd12342/paraspeechcaps .
Authors: David T. Hoffmann, Syed Haseeb Raza, Hanqiu Jiang, Denis Tananaev, Steffen Klingenhoefer, Martin Meinke
Abstract: Scene flow estimation is a foundational task for many robotic applications, including robust dynamic object detection, automatic labeling, and sensor synchronization. Two types of approaches to the problem have evolved: 1) Supervised and 2) optimization-based methods. Supervised methods are fast during inference and achieve high-quality results, however, they are limited by the need for large amounts of labeled training data and are susceptible to domain gaps. In contrast, unsupervised test-time optimization methods do not face the problem of domain gaps but usually suffer from substantial runtime, exhibit artifacts, or fail to converge to the right solution. In this work, we mitigate several limitations of existing optimization-based methods. To this end, we 1) introduce a simple voxel grid-based model that improves over the standard MLP-based formulation in multiple dimensions and 2) introduce a new multiframe loss formulation. 3) We combine both contributions in our new method, termed Floxels. On the Argoverse 2 benchmark, Floxels is surpassed only by EulerFlow among unsupervised methods while achieving comparable performance at a fraction of the computational cost. Floxels achieves a massive speedup of more than ~60 - 140x over EulerFlow, reducing the runtime from a day to 10 minutes per sequence. Over the faster but low-quality baseline, NSFP, Floxels achieves a speedup of ~14x.
Authors: Ritwik Gupta, Rodolfo Corona, Jiaxin Ge, Eric Wang, Dan Klein, Trevor Darrell, David M. Chan
Abstract: Large language models (LLMs) exhibit the ability to generalize given few-shot examples in their input prompt, an emergent capability known as in-context learning (ICL). We investigate whether LLMs utilize ICL to perform structured reasoning in ways that are consistent with a Bayesian framework or rely on pattern matching. Using a controlled setting of biased coin flips, we find that: (1) LLMs often possess biased priors, causing initial divergence in zero-shot settings, (2) in-context evidence outweighs explicit bias instructions, (3) LLMs broadly follow Bayesian posterior updates, with deviations primarily due to miscalibrated priors rather than flawed updates, and (4) attention magnitude has negligible effect on Bayesian inference. With sufficient demonstrations of biased coin flips via ICL, LLMs update their priors in a Bayesian manner.
Authors: Zhuo Chen, Oriol Mayn\'e i Comas, Zhuotao Jin, Di Luo, Marin Solja\v{c}i\'c
Abstract: We rigorously establish a bipartite mutual information scaling law in natural language that governs long-range dependencies. This scaling law, which we show is distinct from and scales independently of the conventional two-point mutual information, is the key to understanding long-context language modeling. Using this scaling law, we formulate the Long-context Language Modeling (L$^2$M) condition, which relates a model's capacity for effective long context length modeling to the scaling of its latent state size for storing past information. Our results are validated through experiments on both transformers and state space models. This work establishes a theoretical foundation that guides the development of large language models toward longer context lengths.
Authors: Shenhao Wang, Baichuan Mo, Yunhan Zheng, Stephane Hess, Jinhua Zhao
Abstract: Numerous studies have compared machine learning (ML) and discrete choice models (DCMs) in predicting travel demand. However, these studies often lack generalizability as they compare models deterministically without considering contextual variations. To address this limitation, our study develops an empirical benchmark by designing a tournament model, thus efficiently summarizing a large number of experiments, quantifying the randomness in model comparisons, and using formal statistical tests to differentiate between the model and contextual effects. This benchmark study compares two large-scale data sources: a database compiled from literature review summarizing 136 experiments from 35 studies, and our own experiment data, encompassing a total of 6,970 experiments from 105 models and 12 model families. This benchmark study yields two key findings. Firstly, many ML models, particularly the ensemble methods and deep learning, statistically outperform the DCM family (i.e., multinomial, nested, and mixed logit models). However, this study also highlights the crucial role of the contextual factors (i.e., data sources, inputs and choice categories), which can explain models' predictive performance more effectively than the differences in model types alone. Model performance varies significantly with data sources, improving with larger sample sizes and lower dimensional alternative sets. After controlling all the model and contextual factors, significant randomness still remains, implying inherent uncertainty in such model comparisons. Overall, we suggest that future researchers shift more focus from context-specific model comparisons towards examining model transferability across contexts and characterizing the inherent uncertainty in ML, thus creating more robust and generalizable next-generation travel demand models.
Authors: Brandon Amos
Abstract: Optimization is a ubiquitous modeling tool and is often deployed in settings which repeatedly solve similar instances of the same problem. Amortized optimization methods use learning to predict the solutions to problems in these settings, exploiting the shared structure between similar problem instances. These methods have been crucial in variational inference and reinforcement learning and are capable of solving optimization problems many orders of magnitudes times faster than traditional optimization methods that do not use amortization. This tutorial presents an introduction to the amortized optimization foundations behind these advancements and overviews their applications in variational inference, sparse coding, gradient-based meta-learning, control, reinforcement learning, convex optimization, optimal transport, and deep equilibrium networks. The source code for this tutorial is available at https://github.com/facebookresearch/amortized-optimization-tutorial.
URLs: https://github.com/facebookresearch/amortized-optimization-tutorial.
Authors: Xihong Yang, Yiqi Wang, Yue Liu, Yi Wen, Lingyuan Meng, Sihang Zhou, Xinwang Liu, En Zhu
Abstract: Graph Neural Networks (GNNs) have achieved promising performance in semi-supervised node classification in recent years. However, the problem of insufficient supervision, together with representation collapse, largely limits the performance of the GNNs in this field. To alleviate the collapse of node representations in semi-supervised scenario, we propose a novel graph contrastive learning method, termed Mixed Graph Contrastive Network (MGCN). In our method, we improve the discriminative capability of the latent embeddings by an interpolation-based augmentation strategy and a correlation reduction mechanism. Specifically, we first conduct the interpolation-based augmentation in the latent space and then force the prediction model to change linearly between samples. Second, we enable the learned network to tell apart samples across two interpolation-perturbed views through forcing the correlation matrix across views to approximate an identity matrix. By combining the two settings, we extract rich supervision information from both the abundant unlabeled nodes and the rare yet valuable labeled nodes for discriminative representation learning. Extensive experimental results on six datasets demonstrate the effectiveness and the generality of MGCN compared to the existing state-of-the-art methods. The code of MGCN is available at https://github.com/xihongyang1999/MGCN on Github.
Authors: Weijie Su
Abstract: Machine learning (ML) and artificial intelligence (AI) conferences including NeurIPS and ICML have experienced a significant decline in peer review quality in recent years. To address this growing challenge, we introduce the Isotonic Mechanism, a computationally efficient approach to enhancing the accuracy of noisy review scores by incorporating authors' private assessments of their submissions. Under this mechanism, authors with multiple submissions are required to rank their papers in descending order of perceived quality. Subsequently, the raw review scores are calibrated based on this ranking to produce adjusted scores. We prove that authors are incentivized to truthfully report their rankings because doing so maximizes their expected utility, modeled as an additive convex function over the adjusted scores. Moreover, the adjusted scores are shown to be more accurate than the raw scores, with improvements being particularly significant when the noise level is high and the author has many submissions -- a scenario increasingly prevalent at large-scale ML/AI conferences. We further investigate whether submission quality information beyond a simple ranking can be truthfully elicited from authors. We establish that a necessary condition for truthful elicitation is that the mechanism be based on pairwise comparisons of the author's submissions. This result underscores the optimality of the Isotonic Mechanism, as it elicits the most fine-grained truthful information among all mechanisms we consider. We then present several extensions, including a demonstration that the mechanism maintains truthfulness even when authors have only partial rather than complete information about their submission quality. Finally, we discuss future research directions, focusing on the practical implementation of the mechanism and the further development of a theoretical framework inspired by our mechanism.
Authors: Joar Skalse, Nikolaus H. R. Howe, Dmitrii Krasheninnikov, David Krueger
Abstract: We provide the first formal definition of reward hacking, a phenomenon where optimizing an imperfect proxy reward function leads to poor performance according to the true reward function. We say that a proxy is unhackable if increasing the expected proxy return can never decrease the expected true return. Intuitively, it might be possible to create an unhackable proxy by leaving some terms out of the reward function (making it "narrower") or overlooking fine-grained distinctions between roughly equivalent outcomes, but we show this is usually not the case. A key insight is that the linearity of reward (in state-action visit counts) makes unhackability a very strong condition. In particular, for the set of all stochastic policies, two reward functions can only be unhackable if one of them is constant. We thus turn our attention to deterministic policies and finite sets of stochastic policies, where non-trivial unhackable pairs always exist, and establish necessary and sufficient conditions for the existence of simplifications, an important special case of unhackability. Our results reveal a tension between using reward functions to specify narrow tasks and aligning AI systems with human values.
Authors: Elad Hazan, Karan Singh
Abstract: This text presents an introduction to an emerging paradigm in control of dynamical systems and differentiable reinforcement learning called online nonstochastic control. The new approach applies techniques from online convex optimization and convex relaxations to obtain new methods with provable guarantees for classical settings in optimal and robust control. The primary distinction between online nonstochastic control and other frameworks is the objective. In optimal control, robust control, and other control methodologies that assume stochastic noise, the goal is to perform comparably to an offline optimal strategy. In online nonstochastic control, both the cost functions as well as the perturbations from the assumed dynamical model are chosen by an adversary. Thus the optimal policy is not defined a priori. Rather, the target is to attain low regret against the best policy in hindsight from a benchmark class of policies. This objective suggests the use of the decision making framework of online convex optimization as an algorithmic methodology. The resulting methods are based on iterative mathematical optimization algorithms, and are accompanied by finite-time regret and computational complexity guarantees.
Authors: Rickard Br\"uel-Gabrielsson, Tongzhou Wang, Manel Baradad, Justin Solomon
Abstract: Despite dropout's ubiquity in machine learning, its effectiveness as a form of data augmentation remains under-explored. We address two key questions: (i) When is dropout effective as an augmentation strategy? (ii) Is dropout uniquely effective under these conditions? To explore these questions, we propose Deep Augmentation, a network- and modality-agnostic method that applies dropout or PCA transformations to targeted layers in neural networks. Through extensive experiments on contrastive learning tasks in NLP, computer vision, and graph learning, we find that uniformly applying dropout across layers does not consistently improve performance. Instead, dropout proves most beneficial in deeper layers and can be matched by alternative augmentations (e.g., PCA). We also show that a stop-gradient operation is critical for ensuring dropout functions effectively as an augmentation, and that performance trends invert when moving from contrastive tasks to supervised tasks. Our analysis suggests that Deep Augmentation helps mitigate inter-layer co-adaptation -- a notable issue in self-supervised learning due to the absence of labeled data. Drawing on these insights, we outline a procedure for selecting the optimal augmentation layer and demonstrate that Deep Augmentation can outperform traditional input-level augmentations. This simple yet powerful approach can be seamlessly integrated into a wide range of architectures and modalities, yielding notable gains in both performance and generalization.
Authors: Yani Zhang, Helmut B\"olcskei
Abstract: We propose a new perspective on deep ReLU networks, namely as circuit counterparts of Lukasiewicz infinite-valued logic -- a many-valued (MV) generalization of Boolean logic. An algorithm for extracting formulae in MV logic from deep ReLU networks is presented. As the algorithm applies to networks with general, in particular also real-valued, weights, it can be used to extract logical formulae from deep ReLU networks trained on data.
Authors: Huy Q. Le, Chu Myaet Thwal, Yu Qiao, Ye Lin Tun, Minh N. H. Nguyen, Eui-Nam Huh, Choong Seon Hong
Abstract: Multimodal federated learning (MFL) has emerged as a decentralized machine learning paradigm, allowing multiple clients with different modalities to collaborate on training a global model across diverse data sources without sharing their private data. However, challenges, such as data heterogeneity and severely missing modalities, pose crucial hindrances to the robustness of MFL, significantly impacting the performance of global model. The occurrence of missing modalities in real-world applications, such as autonomous driving, often arises from factors like sensor failures, leading knowledge gaps during the training process. Specifically, the absence of a modality introduces misalignment during the local training phase, stemming from zero-filling in the case of clients with missing modalities. Consequently, achieving robust generalization in global model becomes imperative, especially when dealing with clients that have incomplete data. In this paper, we propose $\textbf{Multimodal Federated Cross Prototype Learning (MFCPL)}$, a novel approach for MFL under severely missing modalities. Our MFCPL leverages the complete prototypes to provide diverse modality knowledge in modality-shared level with the cross-modal regularization and modality-specific level with cross-modal contrastive mechanism. Additionally, our approach introduces the cross-modal alignment to provide regularization for modality-specific features, thereby enhancing the overall performance, particularly in scenarios involving severely missing modalities. Through extensive experiments on three multimodal datasets, we demonstrate the effectiveness of MFCPL in mitigating the challenges of data heterogeneity and severely missing modalities while improving the overall performance and robustness of MFL.
Authors: Antonio Emanuele Cin\`a, Francesco Villani, Maura Pintor, Lea Sch\"onherr, Battista Biggio, Marcello Pelillo
Abstract: Evaluating the adversarial robustness of deep networks to gradient-based attacks is challenging. While most attacks consider $\ell_2$- and $\ell_\infty$-norm constraints to craft input perturbations, only a few investigate sparse $\ell_1$- and $\ell_0$-norm attacks. In particular, $\ell_0$-norm attacks remain the least studied due to the inherent complexity of optimizing over a non-convex and non-differentiable constraint. However, evaluating adversarial robustness under these attacks could reveal weaknesses otherwise left untested with more conventional $\ell_2$- and $\ell_\infty$-norm attacks. In this work, we propose a novel $\ell_0$-norm attack, called $\sigma$-zero, which leverages a differentiable approximation of the $\ell_0$ norm to facilitate gradient-based optimization, and an adaptive projection operator to dynamically adjust the trade-off between loss minimization and perturbation sparsity. Extensive evaluations using MNIST, CIFAR10, and ImageNet datasets, involving robust and non-robust models, show that $\sigma$\texttt{-zero} finds minimum $\ell_0$-norm adversarial examples without requiring any time-consuming hyperparameter tuning, and that it outperforms all competing sparse attacks in terms of success rate, perturbation size, and efficiency.
Authors: Yu-Guan Hsieh, James Thornton, Eugene Ndiaye, Michal Klein, Marco Cuturi, Pierre Ablin
Abstract: Beyond minimizing a single training loss, many deep learning estimation pipelines rely on an auxiliary objective to quantify and encourage desirable properties of the model (e.g. performance on another dataset, robustness, agreement with a prior). Although the simplest approach to incorporating an auxiliary loss is to sum it with the training loss as a regularizer, recent works have shown that one can improve performance by blending the gradients beyond a simple sum; this is known as gradient surgery. We cast the problem as a constrained minimization problem where the auxiliary objective is minimized among the set of minimizers of the training loss. To solve this bilevel problem, we follow a parameter update direction that combines the training loss gradient and the orthogonal projection of the auxiliary gradient to the training gradient. In a setting where gradients come from mini-batches, we explain how, using a moving average of the training loss gradients, we can carefully maintain this critical orthogonality property. We demonstrate that our method, Bloop, can lead to much better performances on NLP and vision experiments than other gradient surgery methods without EMA.
Authors: Shahryar Zehtabi, Dong-Jun Han, Rohit Parasnis, Seyyedali Hosseinalipour, Christopher G. Brinton
Abstract: Decentralized federated learning (DFL) captures FL settings where both (i) model updates and (ii) model aggregations are exclusively carried out by the clients without a central server. Existing DFL works have mostly focused on settings where clients conduct a fixed number of local updates between local model exchanges, overlooking heterogeneity and dynamics in communication and computation capabilities. In this work, we propose Decentralized Sporadic Federated Learning ($\texttt{DSpodFL}$), a DFL methodology built on a generalized notion of $\textit{sporadicity}$ in both local gradient and aggregation processes. $\texttt{DSpodFL}$ subsumes many existing decentralized optimization methods under a unified algorithmic framework by modeling the per-iteration (i) occurrence of gradient descent at each client and (ii) exchange of models between client pairs as arbitrary indicator random variables, thus capturing $\textit{heterogeneous and time-varying}$ computation/communication scenarios. We analytically characterize the convergence behavior of $\texttt{DSpodFL}$ for both convex and non-convex models and for both constant and diminishing learning rates, under mild assumptions on the communication graph connectivity, data heterogeneity across clients, and gradient noises. We show how our bounds recover existing results from decentralized gradient descent as special cases. Experiments demonstrate that $\texttt{DSpodFL}$ consistently achieves improved training speeds compared with baselines under various system settings.
Authors: Pierre Marion, Anna Korba, Peter Bartlett, Mathieu Blondel, Valentin De Bortoli, Arnaud Doucet, Felipe Llinares-L\'opez, Courtney Paquette, Quentin Berthet
Abstract: We present a new algorithm to optimize distributions defined implicitly by parameterized stochastic diffusions. Doing so allows us to modify the outcome distribution of sampling processes by optimizing over their parameters. We introduce a general framework for first-order optimization of these processes, that performs jointly, in a single loop, optimization and sampling steps. This approach is inspired by recent advances in bilevel optimization and automatic implicit differentiation, leveraging the point of view of sampling as optimization over the space of probability distributions. We provide theoretical guarantees on the performance of our method, as well as experimental results demonstrating its effectiveness. We apply it to training energy-based models and finetuning denoising diffusions.
Authors: Yuxiao Wen, Arthur Jacot
Abstract: We describe the emergence of a Convolution Bottleneck (CBN) structure in CNNs, where the network uses its first few layers to transform the input representation into a representation that is supported only along a few frequencies and channels, before using the last few layers to map back to the outputs. We define the CBN rank, which describes the number and type of frequencies that are kept inside the bottleneck, and partially prove that the parameter norm required to represent a function $f$ scales as depth times the CBN rank $f$. We also show that the parameter norm depends at next order on the regularity of $f$. We show that any network with almost optimal parameter norm will exhibit a CBN structure in both the weights and - under the assumption that the network is stable under large learning rate - the activations, which motivates the common practice of down-sampling; and we verify that the CBN results still hold with down-sampling. Finally we use the CBN structure to interpret the functions learned by CNNs on a number of tasks.
Authors: Achraf Azize, Debabrota Basu
Abstract: In a Membership Inference (MI) game, an attacker tries to infer whether a target point was included or not in the input of an algorithm. Existing works show that some target points are easier to identify, while others are harder. This paper explains the target-dependent hardness of membership attacks by studying the powers of the optimal attacks in a fixed-target MI game. We characterise the optimal advantage and trade-off functions of attacks against the empirical mean in terms of the Mahalanobis distance between the target point and the data-generating distribution. We further derive the impacts of two privacy defences, i.e. adding Gaussian noise and sub-sampling, and that of target misspecification on optimal attacks. As by-products of our novel analysis of the Likelihood Ratio (LR) test, we provide a new covariance attack which generalises and improves the scalar product attack. Also, we propose a new optimal canary-choosing strategy for auditing privacy in the white-box federated learning setting. Our experiments validate that the Mahalanobis score explains the hardness of fixed-target MI games.
Authors: Laura Manduchi, Kushagra Pandey, Clara Meister, Robert Bamler, Ryan Cotterell, Sina D\"aubener, Sophie Fellenz, Asja Fischer, Thomas G\"artner, Matthias Kirchler, Marius Kloft, Yingzhen Li, Christoph Lippert, Gerard de Melo, Eric Nalisnick, Bj\"orn Ommer, Rajesh Ranganath, Maja Rudolph, Karen Ullrich, Guy Van den Broeck, Julia E Vogt, Yixin Wang, Florian Wenzel, Frank Wood, Stephan Mandt, Vincent Fortuin
Abstract: The field of deep generative modeling has grown rapidly in the last few years. With the availability of massive amounts of training data coupled with advances in scalable unsupervised learning paradigms, recent large-scale generative models show tremendous promise in synthesizing high-resolution images and text, as well as structured data such as videos and molecules. However, we argue that current large-scale generative AI models exhibit several fundamental shortcomings that hinder their widespread adoption across domains. In this work, our objective is to identify these issues and highlight key unresolved challenges in modern generative AI paradigms that should be addressed to further enhance their capabilities, versatility, and reliability. By identifying these challenges, we aim to provide researchers with insights for exploring fruitful research directions, thus fostering the development of more robust and accessible generative AI solutions.
Authors: Alhassan Mumuni, Fuseini Mumuni
Abstract: Data augmentation is arguably the most important regularization technique commonly used to improve generalization performance of machine learning models. It primarily involves the application of appropriate data transformation operations to create new data samples with desired properties. Despite its effectiveness, the process is often challenging because of the time-consuming trial and error procedures for creating and testing different candidate augmentations and their hyperparameters manually. State-of-the-art approaches are increasingly relying on automated machine learning (AutoML) principles. This work presents a comprehensive survey of AutoML-based data augmentation techniques. We discuss various approaches for accomplishing data augmentation with AutoML, including data manipulation, data integration and data synthesis techniques. The focus of this work is on image data augmentation methods. Nonetheless, we cover other data modalities, especially in cases where the specific data augmentations techniques being discussed are more suitable for these other modalities. For instance, since automated data integration methods are more suitable for tabular data, we cover tabular data in the discussion of data integration methods. The work also presents extensive discussion of techniques for accomplishing each of the major subtasks of the image data augmentation process: search space design, hyperparameter optimization and model evaluation. Finally, we carried out an extensive comparison and analysis of the performance of automated data augmentation techniques and state-of-the-art methods based on classical augmentation approaches. The results show that AutoML methods for data augmentation currently outperform state-of-the-art techniques based on conventional approaches.
Authors: William Merrill, Jackson Petty, Ashish Sabharwal
Abstract: State-space models (SSMs) have emerged as a potential alternative architecture for building large language models (LLMs) compared to the previously ubiquitous transformer architecture. One theoretical weakness of transformers is that they cannot express certain kinds of sequential computation and state tracking (Merrill & Sabharwal, 2023), which SSMs are explicitly designed to address via their close architectural similarity to recurrent neural networks (RNNs). But do SSMs truly have an advantage (over transformers) in expressive power for state tracking? Surprisingly, the answer is no. Our analysis reveals that the expressive power of SSMs is limited very similarly to transformers: SSMs cannot express computation outside the complexity class $\mathsf{TC}^0$. In particular, this means they cannot solve simple state-tracking problems like permutation composition. It follows that SSMs are provably unable to accurately track chess moves with certain notation, evaluate code, or track entities in a long narrative. To supplement our formal analysis, we report experiments showing that Mamba-style SSMs indeed struggle with state tracking. Thus, despite its recurrent formulation, the "state" in an SSM is an illusion: SSMs have similar expressiveness limitations to non-recurrent models like transformers, which may fundamentally limit their ability to solve real-world state-tracking problems.
Authors: Sanath Kumar Krishnamurthy, Anna Lyubarskaja, Emma Brunskill, Susan Athey
Abstract: Constructing confidence intervals that are simultaneously valid across a class of estimates is central for tasks such as multiple mean estimation, bounding generalization error in machine learning, and adaptive experimental design. We frame this as an "error estimation problem," where the goal is to determine a high-probability upper bound on the maximum error for a class of estimates. We propose an entirely data-driven approach that derives such bounds for both finite and infinite class settings, naturally adapting to a potentially unknown correlation structure of random errors. Notably, our method does not require class complexity as an input, overcoming a major limitation of existing approaches such as union bounding and bounds based on Talagrand's inequality. In this paper, we present our simple yet general solution and demonstrate its flexibility through applications ranging from constructing multiple simultaneously valid confidence intervals to optimizing exploration in contextual bandit algorithms.
Authors: Stephen Pasteris, Chris Hicks, Vasilios Mavroudis, Mark Herbster
Abstract: We consider the classic problem of online convex optimisation. Whereas the notion of static regret is relevant for stationary problems, the notion of switching regret is more appropriate for non-stationary problems. A switching regret is defined relative to any segmentation of the trial sequence, and is equal to the sum of the static regrets of each segment. In this paper we show that, perhaps surprisingly, we can achieve the asymptotically optimal switching regret on every possible segmentation simultaneously. Our algorithm for doing so is very efficient: having a space and per-trial time complexity that is logarithmic in the time-horizon. Our algorithm also obtains novel bounds on its dynamic regret: being adaptive to variations in the rate of change of the comparator sequence.
Authors: Utsav Singh, Pramit Bhattacharyya, Vinay P. Namboodiri
Abstract: Developing interactive systems that utilize natural language instructions to solve complex robotic control tasks has long been a goal of the robotics community. While Large Language Models (LLMs) excel at logical reasoning, in-context learning, and code generation, translating high-level instructions into low-level robotic actions still remains challenging. Furthermore, solving such tasks often requires acquiring policies to execute diverse subtasks and integrating them to achieve the final objective. Hierarchical Reinforcement Learning (HRL) offers a promising solution for solving such tasks by enabling temporal abstraction and improved exploration. However, HRL suffers from non-stationarity caused by the changing lower-level behaviour, which hinders effective policy learning. We propose LGR2, a novel HRL framework that mitigates non-stationarity in HRL by using language-guided higher-level rewards that remain unaffected by the changing lower-level policy behaviour. To analyze the efficacy of our approach, we perform empirical analysis to demonstrate that LGR2 effectively mitigates non-stationarity in HRL and attains success rates exceeding 70% in challenging, sparsely-rewarded robotic navigation and manipulation environments, where other baselines typically fail to show significant progress. Finally, we perform real-world robotic experiments on complex tasks and demonstrate that LGR2 consistently outperforms the baselines.
Authors: Roman Bresson, Giannis Nikolentzos, George Panagopoulos, Michail Chatzianastasis, Jun Pang, Michalis Vazirgiannis
Abstract: In recent years, Graph Neural Networks (GNNs) have become the de facto tool for learning node and graph representations. Most GNNs typically consist of a sequence of neighborhood aggregation (a.k.a., message-passing) layers, within which the representation of each node is updated based on those of its neighbors. The most expressive message-passing GNNs can be obtained through the use of the sum aggregator and of MLPs for feature transformation, thanks to their universal approximation capabilities. However, the limitations of MLPs recently motivated the introduction of another family of universal approximators, called Kolmogorov-Arnold Networks (KANs) which rely on a different representation theorem. In this work, we compare the performance of KANs against that of MLPs on graph learning tasks. We implement three new KAN-based GNN layers, inspired respectively by the GCN, GAT and GIN layers. We evaluate two different implementations of KANs using two distinct base families of functions, namely B-splines and radial basis functions. We perform extensive experiments on node classification, link prediction, graph classification and graph regression datasets. Our results indicate that KANs are on-par with or better than MLPs on all tasks studied in this paper. We also show that the size and training speed of RBF-based KANs is only marginally higher than for MLPs, making them viable alternatives. Code available at https://github.com/RomanBresson/KAGNN.
Authors: Jeremy Carleton, Prathik Vijaykumar, Divyanshu Saxena, Dheeraj Narasimha, Srinivas Shakkottai, Aditya Akella
Abstract: We address the challenge of zeroth-order online convex optimization where the objective function's gradient exhibits sparsity, indicating that only a small number of dimensions possess non-zero gradients. Our aim is to leverage this sparsity to obtain useful estimates of the objective function's gradient even when the only information available is a limited number of function samples. Our motivation stems from the optimization of large-scale queueing networks that process time-sensitive jobs. Here, a job must be processed by potentially many queues in sequence to produce an output, and the service time at any queue is a function of the resources allocated to that queue. Since resources are costly, the end-to-end latency for jobs must be balanced with the overall cost of the resources used. While the number of queues is substantial, the latency function primarily reacts to resource changes in only a few, rendering the gradient sparse. We tackle this problem by introducing the Compressive Online Gradient Optimization framework which allows compressive sensing methods previously applied to stochastic optimization to achieve regret bounds with an optimal dependence on the time horizon without the full problem dimension appearing in the bound. For specific algorithms, we reduce the samples required per gradient estimate to scale with the gradient's sparsity factor rather than its full dimensionality. Numerical simulations and real-world microservices benchmarks demonstrate CONGO's superiority over gradient descent approaches that do not account for sparsity.
Authors: Mengjie Zhao, Cees Taal, Stephan Baggerohr, Olga Fink
Abstract: Real-time condition monitoring is crucial for the reliable and efficient operation of complex systems. However, relying solely on physical sensors can be limited due to their cost, placement constraints, or inability to directly measure certain critical parameters. Virtual sensing addresses these limitations by leveraging readily available sensor data and system knowledge to estimate inaccessible parameters or infer system states. The increasing complexity of industrial systems necessitates deployments of sensors with diverse modalities to provide a comprehensive understanding of system states. These sensors capture data at varying frequencies to monitor both rapid and slowly varying system dynamics, as well as local and global state evolutions of the systems. This leads to heterogeneous temporal dynamics, which, particularly under varying operational end environmental conditions, pose a significant challenge for accurate virtual sensing. To address this, we propose a Heterogeneous Temporal Graph Neural Network (HTGNN) framework. HTGNN explicitly models signals from diverse sensors and integrates operating conditions into the model architecture. We evaluate HTGNN using two newly released datasets: a bearing dataset with diverse load conditions for bearing load prediction and a year-long simulated dataset for predicting bridge live loads. Our results demonstrate that HTGNN significantly outperforms established baseline methods in both tasks, particularly under highly varying operating conditions. These results highlight HTGNN's potential as a robust and accurate virtual sensing approach for complex systems, paving the way for improved monitoring, predictive maintenance, and enhanced system performance. Our code and data are available under https://github.com/EPFL-IMOS/htgnn.
Authors: Yupei Yang, Biwei Huang, Fan Feng, Xinyue Wang, Shikui Tu, Lei Xu
Abstract: General intelligence requires quick adaption across tasks. While existing reinforcement learning (RL) methods have made progress in generalization, they typically assume only distribution changes between source and target domains. In this paper, we explore a wider range of scenarios where not only the distribution but also the environment spaces may change. For example, in the CoinRun environment, we train agents from easy levels and generalize them to difficulty levels where there could be new enemies that have never occurred before. To address this challenging setting, we introduce a causality-guided self-adaptive representation-based approach, called CSR, that equips the agent to generalize effectively across tasks with evolving dynamics. Specifically, we employ causal representation learning to characterize the latent causal variables within the RL system. Such compact causal representations uncover the structural relationships among variables, enabling the agent to autonomously determine whether changes in the environment stem from distribution shifts or variations in space, and to precisely locate these changes. We then devise a three-step strategy to fine-tune the causal model under different scenarios accordingly. Empirical experiments show that CSR efficiently adapts to the target domains with only a few samples and outperforms state-of-the-art baselines on a wide range of scenarios, including our simulated environments, CartPole, CoinRun and Atari games.
Authors: Stefano Calzavara, Lorenzo Cazzaro, Massimo Vettori
Abstract: We present Timber, the first white-box poisoning attack targeting decision trees. Timber is based on a greedy attack strategy that leverages sub-tree retraining to efficiently estimate the damage caused by poisoning a given training instance. The attack relies on a tree annotation procedure, which enables the sorting of training instances so that they are processed in increasing order of the computational cost of sub-tree retraining. This sorting yields a variant of Timber that supports an early stopping criterion, designed to make poisoning attacks more efficient and feasible on larger datasets. We also discuss an extension of Timber to traditional random forest models, which is valuable since decision trees are typically combined into ensembles to improve their predictive power. Our experimental evaluation on public datasets demonstrates that our attacks outperform existing baselines in terms of effectiveness, efficiency, or both. Moreover, we show that two representative defenses can mitigate the effect of our attacks, but fail to effectively thwart them.
Authors: Zhilin Wang, Alexander Bukharin, Olivier Delalleau, Daniel Egert, Gerald Shen, Jiaqi Zeng, Oleksii Kuchaiev, Yi Dong
Abstract: Reward models are critical for aligning models to follow instructions, and are typically trained following one of two popular paradigms: Bradley-Terry style or Regression style. However, there is a lack of evidence that either approach is better than the other, when adequately matched for data. This is primarily because these approaches require data collected in different (but incompatible) formats, meaning that adequately matched data is not available in existing public datasets. To tackle this problem, we release preference annotations (designed for Bradley-Terry training) to complement existing ratings (designed for Regression style training) in the HelpSteer2 dataset. To improve data interpretability, preference annotations are accompanied with human-written justifications. Using this data, we conduct the first head-to-head comparison of Bradley-Terry and Regression models when adequately matched for data. Based on insights derived from such a comparison, we propose a novel approach to combine Bradley-Terry and Regression reward modeling. A Llama-3.1-70B-Instruct model tuned with this approach scores 94.1 on RewardBench, emerging top of more than 140 reward models as of 1 Oct 2024. This reward model can then be used with REINFORCE algorithm (RLHF) to align an Instruct model to reach 85.0 on Arena Hard, which is No. 1 as of 1 Oct 2024. We open-source this dataset (CC-BY-4.0 license) at https://huggingface.co/datasets/nvidia/HelpSteer2#preferences-new -- 1-oct-2024 and openly release the trained Reward and Instruct models at https://huggingface.co/nvidia/Llama-3.1-Nemotron-70B-Reward and https://huggingface.co/nvidia/Llama-3.1-Nemotron-70B-Instruct
URLs: https://huggingface.co/datasets/nvidia/HelpSteer2, https://huggingface.co/nvidia/Llama-3.1-Nemotron-70B-Reward, https://huggingface.co/nvidia/Llama-3.1-Nemotron-70B-Instruct
Authors: Yongxing Zhang, Donglin Yang, Renjie Liao
Abstract: Finite symmetric groups $S_n$ are essential in fields such as combinatorics, physics, and chemistry. However, learning a probability distribution over $S_n$ poses significant challenges due to its intractable size and discrete nature. In this paper, we introduce SymmetricDiffusers, a novel discrete diffusion model that simplifies the task of learning a complicated distribution over $S_n$ by decomposing it into learning simpler transitions of the reverse diffusion using deep neural networks. We identify the riffle shuffle as an effective forward transition and provide empirical guidelines for selecting the diffusion length based on the theory of random walks on finite groups. Additionally, we propose a generalized Plackett-Luce (PL) distribution for the reverse transition, which is provably more expressive than the PL distribution. We further introduce a theoretically grounded "denoising schedule" to improve sampling and learning efficiency. Extensive experiments show that our model achieves state-of-the-art or comparable performances on solving tasks including sorting 4-digit MNIST images, jigsaw puzzles, and traveling salesman problems. Our code is released at https://github.com/DSL-Lab/SymmetricDiffusers.
Authors: Abbas Abdolmaleki, Bilal Piot, Bobak Shahriari, Jost Tobias Springenberg, Tim Hertweck, Rishabh Joshi, Junhyuk Oh, Michael Bloesch, Thomas Lampe, Nicolas Heess, Jonas Buchli, Martin Riedmiller
Abstract: Existing preference optimization methods often assume scenarios where paired preference feedback (preferred/positive vs. dis-preferred/negative examples) is available. This requirement limits their applicability in scenarios where only unpaired feedback--for example, either positive or negative--is available. To address this, we introduce a novel approach that decouples learning from positive and negative feedback. This decoupling enables control over the influence of each feedback type and, importantly, allows learning even when only one feedback type is present. A key contribution is demonstrating stable learning from negative feedback alone, a capability not well-addressed by current methods. Our approach builds upon the probabilistic framework introduced in (Dayan and Hinton, 1997), which uses expectation-maximization (EM) to directly optimize the probability of positive outcomes (as opposed to classic expected reward maximization). We address a key limitation in current EM-based methods: they solely maximize the likelihood of positive examples, while neglecting negative ones. We show how to extend EM algorithms to explicitly incorporate negative examples, leading to a theoretically grounded algorithm that offers an intuitive and versatile way to learn from both positive and negative feedback. We evaluate our approach for training language models based on human feedback as well as training policies for sequential decision-making problems, where learned value functions are available.
Authors: Ayano Hiranaka, Shang-Fu Chen, Chieh-Hsin Lai, Dongjun Kim, Naoki Murata, Takashi Shibuya, Wei-Hsiang Liao, Shao-Hua Sun, Yuki Mitsufuji
Abstract: Controllable generation through Stable Diffusion (SD) fine-tuning aims to improve fidelity, safety, and alignment with human guidance. Existing reinforcement learning from human feedback methods usually rely on predefined heuristic reward functions or pretrained reward models built on large-scale datasets, limiting their applicability to scenarios where collecting such data is costly or difficult. To effectively and efficiently utilize human feedback, we develop a framework, HERO, which leverages online human feedback collected on the fly during model learning. Specifically, HERO features two key mechanisms: (1) Feedback-Aligned Representation Learning, an online training method that captures human feedback and provides informative learning signals for fine-tuning, and (2) Feedback-Guided Image Generation, which involves generating images from SD's refined initialization samples, enabling faster convergence towards the evaluator's intent. We demonstrate that HERO is 4x more efficient in online feedback for body part anomaly correction compared to the best existing method. Additionally, experiments show that HERO can effectively handle tasks like reasoning, counting, personalization, and reducing NSFW content with only 0.5K online feedback.
Authors: Yonatan Sverdlov, Ido Springer, Nadav Dym
Abstract: This paper explores the characterization of equivariant linear layers for representations of permutations and related groups. Unlike traditional approaches, which address these problems using parameter-sharing, we consider an alternative methodology based on irreducible representations and Schur's lemma. Using this methodology, we obtain an alternative derivation for existing models like DeepSets, 2-IGN graph equivariant networks, and Deep Weight Space (DWS) networks. The derivation for DWS networks is significantly simpler than that of previous results. Next, we extend our approach to unaligned symmetric sets, where equivariance to the wreath product of groups is required. Previous works have addressed this problem in a rather restrictive setting, in which almost all wreath equivariant layers are Siamese. In contrast, we give a full characterization of layers in this case and show that there is a vast number of additional non-Siamese layers in some settings. We also show empirically that these additional non-Siamese layers can improve performance in tasks like graph anomaly detection, weight space alignment, and learning Wasserstein distances. Our code is available at \href{https://github.com/yonatansverdlov/Irreducible-Representations-of-Deep-Weight-Spaces}{GitHub}.
URLs: https://github.com/yonatansverdlov/Irreducible-Representations-of-Deep-Weight-Spaces
Authors: Michael Zhang, Simran Arora, Rahul Chalamala, Alan Wu, Benjamin Spector, Aaryan Singhal, Krithik Ramesh, Christopher R\'e
Abstract: Recent works show we can linearize large language models (LLMs) -- swapping the quadratic attentions of popular Transformer-based LLMs with subquadratic analogs, such as linear attention -- avoiding the expensive pretraining costs. However, linearizing LLMs often significantly degrades model quality, still requires training over billions of tokens, and remains limited to smaller 1.3B to 7B LLMs. We thus propose Low-rank Linear Conversion via Attention Transfer (LoLCATs), a simple two-step method that improves LLM linearizing quality with orders of magnitudes less memory and compute. We base these steps on two findings. First, we can replace an LLM's softmax attentions with closely-approximating linear attentions, simply by training the linear attentions to match their softmax counterparts with an output MSE loss ("attention transfer"). Then, this enables adjusting for approximation errors and recovering LLM quality simply with low-rank adaptation (LoRA). LoLCATs significantly improves linearizing quality, training efficiency, and scalability. We significantly reduce the linearizing quality gap and produce state-of-the-art subquadratic LLMs from Llama 3 8B and Mistral 7B v0.1, leading to 20+ points of improvement on 5-shot MMLU. Furthermore, LoLCATs does so with only 0.2% of past methods' model parameters and 0.4% of their training tokens. Finally, we apply LoLCATs to create the first linearized 70B and 405B LLMs (50x larger than prior work). When compared with prior approaches under the same compute budgets, LoLCATs significantly improves linearizing quality, closing the gap between linearized and original Llama 3.1 70B and 405B LLMs by 77.8% and 78.1% on 5-shot MMLU.
Authors: Fanqi Yan, Huy Nguyen, Dung Le, Pedram Akbarian, Nhat Ho
Abstract: We conduct the convergence analysis of parameter estimation in the contaminated mixture of experts. This model is motivated from the prompt learning problem where ones utilize prompts, which can be formulated as experts, to fine-tune a large-scale pre-trained model for learning downstream tasks. There are two fundamental challenges emerging from the analysis: (i) the proportion in the mixture of the pre-trained model and the prompt may converge to zero during the training, leading to the prompt vanishing issue; (ii) the algebraic interaction among parameters of the pre-trained model and the prompt can occur via some partial differential equations and decelerate the prompt learning. In response, we introduce a distinguishability condition to control the previous parameter interaction. Additionally, we also investigate various types of expert structure to understand their effects on the convergence behavior of parameter estimation. In each scenario, we provide comprehensive convergence rates of parameter estimation along with the corresponding minimax lower bounds. Finally, we run several numerical experiments to empirically justify our theoretical findings.
Authors: Xingjian Wu, Xiangfei Qiu, Zhengyu Li, Yihang Wang, Jilin Hu, Chenjuan Guo, Hui Xiong, Bin Yang
Abstract: Anomaly detection in multivariate time series is challenging as heterogeneous subsequence anomalies may occur. Reconstruction-based methods, which focus on learning normal patterns in the frequency domain to detect diverse abnormal subsequences, achieve promising results, while still falling short on capturing fine-grained frequency characteristics and channel correlations. To contend with the limitations, we introduce CATCH, a framework based on frequency patching. We propose to patchify the frequency domain into frequency bands, which enhances its ability to capture fine-grained frequency characteristics. To perceive appropriate channel correlations, we propose a Channel Fusion Module (CFM), which features a patch-wise mask generator and a masked-attention mechanism. Driven by a bi-level multi-objective optimization algorithm, the CFM is encouraged to iteratively discover appropriate patch-wise channel correlations, and to cluster relevant channels while isolating adverse effects from irrelevant channels. Extensive experiments on 10 real-world datasets and 12 synthetic datasets demonstrate that CATCH achieves state-of-the-art performance. We make our code and datasets available at https://github.com/decisionintelligence/CATCH.
Authors: Dominik Fuchsgruber, Tim Po\v{s}tuvan, Stephan G\"unnemann, Simon Geisler
Abstract: Many applications in traffic, civil engineering, or electrical engineering revolve around edge-level signals. Such signals can be categorized as inherently directed, for example, the water flow in a pipe network, and undirected, like the diameter of a pipe. Topological methods model edge signals with inherent direction by representing them relative to a so-called orientation assigned to each edge. These approaches can neither model undirected edge signals nor distinguish if an edge itself is directed or undirected. We address these shortcomings by (i) revising the notion of orientation equivariance to enable edge direction-aware topological models, (ii) proposing orientation invariance as an additional requirement to describe signals without inherent direction, and (iii) developing EIGN, an architecture composed of novel direction-aware edge-level graph shift operators, that provably fulfills the aforementioned desiderata. It is the first general-purpose topological GNN for edge-level signals that can model directed and undirected signals while distinguishing between directed and undirected edges. A comprehensive evaluation shows that EIGN outperforms prior work in edge-level tasks, for example, improving in RMSE on flow simulation tasks by up to 23.5%.
Authors: Yufeng Wu, Rohit Bhattacharya
Abstract: A key question in many network studies is whether the observed correlations between units are primarily due to contagion or latent confounding. Here, we study this question using a segregated graph (Shpitser, 2015) representation of these mechanisms, and examine how uncertainty about the true underlying mechanism impacts downstream computation of network causal effects, particularly under full interference -- settings where we only have a single realization of a network and each unit may depend on any other unit in the network. Under certain assumptions about asymptotic growth of the network, we derive likelihood ratio tests that can be used to identify whether different sets of variables -- confounders, treatments, and outcomes -- across units exhibit dependence due to contagion or latent confounding. We then propose network causal effect estimation strategies that provide unbiased and consistent estimates if the dependence mechanisms are either known or correctly inferred using our proposed tests. Together, the proposed methods allow network effect estimation in a wider range of full interference scenarios that have not been considered in prior work. We evaluate the effectiveness of our methods with synthetic data and the validity of our assumptions using real-world networks.
Authors: Hu Wang, Congbo Ma, Ibrahim Almakky, Ian Reid, Gustavo Carneiro, Mohammad Yaqub
Abstract: Model-merging has emerged as a powerful approach in deep learning, capable of enhancing model performance without any training. However, the underlying mechanisms that explain its effectiveness remain largely unexplored. In this paper, we investigate this technique from three novel perspectives to empirically provide deeper insights into why and how weight-averaged model-merging works: (1) we examine the intrinsic patterns captured by the learning of the model weights, through the visualizations of their patterns on several datasets, showing that these weights often encode structured and interpretable patterns and that is the essential why model-merging can work; (2) we mathematically and empirically investigate model ensemble merging strategies based on averaging on weights versus averaging on features, providing detailed analyses across diverse architectures and datasets; and (3) we explore the impact on model-merging prediction stability in terms of changing the parameter magnitude, revealing insights into the way of weight averaging works as regularization by showing the robustness across different parameter scales. Our findings shed light on the "black box" of weight-averaged model-merging, offering valuable insights and practical recommendations that advance the model-merging process. The code is available at https://github.com/billhhh/Rethink-Merge.
Authors: Joel Daniel Andersson, Rasmus Pagh
Abstract: In differential privacy, $\textit{continual observation}$ refers to problems in which we wish to continuously release a function of a dataset that is revealed one element at a time. The challenge is to maintain a good approximation while keeping the combined output over all time steps differentially private. In the special case of $\textit{continual counting}$ we seek to approximate a sum of binary input elements. This problem has received considerable attention lately, in part due to its relevance in implementations of differentially private stochastic gradient descent. $\textit{Factorization mechanisms}$ are the leading approach to continual counting, but the best such mechanisms do not work well in $\textit{streaming}$ settings since they require space proportional to the size of the input. In this paper, we present a simple approach to approximating factorization mechanisms in low space via $\textit{binning}$, where adjacent matrix entries with similar values are changed to be identical in such a way that a matrix-vector product can be maintained in sublinear space. Our approach has provable sublinear space guarantees for a class of lower triangular matrices whose entries are monotonically decreasing away from the diagonal. We show empirically that even with very low space usage we are able to closely match, and sometimes surpass, the performance of asymptotically optimal factorization mechanisms. Recently, and independently of our work, Dvijotham et al. have also suggested an approach to implementing factorization mechanisms in a streaming setting. Their work differs from ours in several respects: It only addresses factorization into $\textit{Toeplitz}$ matrices, only considers $\textit{maximum}$ error, and uses a different technique based on rational function approximation that seems less versatile than our binning approach.
Authors: Zhen Liu, Tim Z. Xiao, Weiyang Liu, Yoshua Bengio, Dinghuai Zhang
Abstract: While one commonly trains large diffusion models by collecting datasets on target downstream tasks, it is often desired to align and finetune pretrained diffusion models with some reward functions that are either designed by experts or learned from small-scale datasets. Existing post-training methods for reward finetuning of diffusion models typically suffer from lack of diversity in generated samples, lack of prior preservation, and/or slow convergence in finetuning. Inspired by recent successes in generative flow networks (GFlowNets), a class of probabilistic models that sample with the unnormalized density of a reward function, we propose a novel GFlowNet method dubbed Nabla-GFlowNet (abbreviated as \methodname), the first GFlowNet method that leverages the rich signal in reward gradients, together with an objective called \graddb plus its variant \resgraddb designed for prior-preserving diffusion finetuning. We show that our proposed method achieves fast yet diversity- and prior-preserving finetuning of Stable Diffusion, a large-scale text-conditioned image diffusion model, on different realistic reward functions.
Authors: Alex F. Spies, William Edwards, Michael I. Ivanitskiy, Adrians Skapars, Tilman R\"auker, Katsumi Inoue, Alessandra Russo, Murray Shanahan
Abstract: Recent studies in interpretability have explored the inner workings of transformer models trained on tasks across various domains, often discovering that these networks naturally develop highly structured representations. When such representations comprehensively reflect the task domain's structure, they are commonly referred to as "World Models" (WMs). In this work, we identify WMs in transformers trained on maze-solving tasks. By using Sparse Autoencoders (SAEs) and analyzing attention patterns, we examine the construction of WMs and demonstrate consistency between SAE feature-based and circuit-based analyses. By subsequently intervening on isolated features to confirm their causal role, we find that it is easier to activate features than to suppress them. Furthermore, we find that models can reason about mazes involving more simultaneously active features than they encountered during training; however, when these same mazes (with greater numbers of connections) are provided to models via input tokens instead, the models fail. Finally, we demonstrate that positional encoding schemes appear to influence how World Models are structured within the model's residual stream.
Authors: Tengxue Zhang, Yang Shu, Xinyang Chen, Yifei Long, Chenjuan Guo, Bin Yang
Abstract: Pre-trained model assessment for transfer learning aims to identify the optimal candidate for the downstream tasks from a model hub, without the need of time-consuming fine-tuning. Existing advanced works mainly focus on analyzing the intrinsic characteristics of the entire features extracted by each pre-trained model or how well such features fit the target labels. This paper proposes a novel perspective for pre-trained model assessment through the Distribution of Spectral Components (DISCO). Through singular value decomposition of features extracted from pre-trained models, we investigate different spectral components and observe that they possess distinct transferability, contributing diversely to the fine-tuning performance. Inspired by this, we propose an assessment method based on the distribution of spectral components which measures the proportions of their corresponding singular values. Pre-trained models with features concentrating on more transferable components are regarded as better choices for transfer learning. We further leverage the labels of downstream data to better estimate the transferability of each spectral component and derive the final assessment criterion. Our proposed method is flexible and can be applied to both classification and regression tasks. We conducted comprehensive experiments across three benchmarks and two tasks including image classification and object detection, demonstrating that our method achieves state-of-the-art performance in choosing proper pre-trained models from the model hub for transfer learning.
Authors: Yuchen Lin, Yong Zhang, Sihan Feng, Hong Zhao
Abstract: Advancing artificial intelligence demands a deeper understanding of the mechanisms underlying deep learning. Here, we propose a straightforward analysis framework based on the dynamics of learning models. Neurons are categorized into two modes based on whether their transformation functions preserve order. This categorization reveals how deep neural networks (DNNs) maximize information extraction by rationally allocating the proportion of neurons in different modes across deep layers. We further introduce the attraction basins of the training samples in both the sample vector space and the weight vector space to characterize the generalization ability of DNNs. This framework allows us to identify optimal depth and width configurations, providing a unified explanation for fundamental DNN behaviors such as the "flat minima effect," "grokking," and double descent phenomena. Our analysis extends to networks with depths up to 100 layers.
Authors: Michael J. Zellinger, Matt Thomson
Abstract: Understanding the reliability of large language models (LLMs) has recently garnered significant attention. Given LLMs' propensity to hallucinate, as well as their high sensitivity to prompt design, it is already challenging to predict the performance of an individual LLM. However, the problem becomes more complex for compound LLM systems such as cascades, where in addition to each model's standalone performance, we must understand how the error rates of different models interact. In this paper, we present a probabilistic model for the joint performance distribution of a sequence of LLMs, which enables a framework for rationally tuning the confidence thresholds of a LLM cascade using continuous optimization. Compared to selecting confidence thresholds using grid search, our parametric Markov-copula model significantly improves runtime scaling with respect to the length of the cascade and the desired resolution of the cost-error curve, turning them from intractable into low-order polynomial. In addition, the optimal thresholds computed using our continuous optimization-based algorithm increasingly outperform those found via grid search as cascade length grows, improving the area under the cost-error curve by 1.9% on average for cascades consisting of at least three models. Overall, our Markov-copula model provides a rational basis for tuning LLM cascade performance and points to the potential of probabilistic methods in analyzing LLM systems.
Authors: Rui Xu, Chao Chen, Yue Sun, Parvathinathan Venkitasubramaniam, Sihong Xie
Abstract: Conformal prediction yields a prediction set with guaranteed $1-\alpha$ coverage of the true target under the i.i.d. assumption, which may not hold and lead to a gap between $1-\alpha$ and the actual coverage. Prior studies bound the gap using total variation distance, which cannot identify the gap changes under distribution shift at a given $\alpha$. Besides, existing methods are mostly limited to covariate shift,while general joint distribution shifts are more common in practice but less researched.In response, we first propose a Wasserstein distance-based upper bound of the coverage gap and analyze the bound using probability measure pushforwards between the shifted joint data and conformal score distributions, enabling a separation of the effect of covariate and concept shifts over the coverage gap. We exploit the separation to design an algorithm based on importance weighting and regularized representation learning (WR-CP) to reduce the Wasserstein bound with a finite-sample error bound.WR-CP achieves a controllable balance between conformal prediction accuracy and efficiency. Experiments on six datasets prove that WR-CP can reduce coverage gaps to $3.2\%$ across different confidence levels and outputs prediction sets 37$\%$ smaller than the worst-case approach on average.
Authors: Zhi Sheng, Yuan Yuan, Jingtao Ding, Yong Li
Abstract: Accurate prediction of mobile traffic, \textit{i.e.,} network traffic from cellular base stations, is crucial for optimizing network performance and supporting urban development. However, the non-stationary nature of mobile traffic, driven by human activity and environmental changes, leads to both regular patterns and abrupt variations. Diffusion models excel in capturing such complex temporal dynamics due to their ability to capture the inherent uncertainties. Most existing approaches prioritize designing novel denoising networks but often neglect the critical role of noise itself, potentially leading to sub-optimal performance. In this paper, we introduce a novel perspective by emphasizing the role of noise in the denoising process. Our analysis reveals that noise fundamentally shapes mobile traffic predictions, exhibiting distinct and consistent patterns. We propose NPDiff, a framework that decomposes noise into \textit{prior} and \textit{residual} components, with the \textit{prior} derived from data dynamics, enhancing the model's ability to capture both regular and abrupt variations. NPDiff can seamlessly integrate with various diffusion-based prediction models, delivering predictions that are effective, efficient, and robust. Extensive experiments demonstrate that it achieves superior performance with an improvement over 30\%, offering a new perspective on leveraging diffusion models in this domain.
Authors: Vivek Bharadwaj, Austin Glover, Aydin Buluc, James Demmel
Abstract: Rotation equivariant graph neural networks, i.e., networks designed to guarantee certain geometric relations between their inputs and outputs, yield state-of-the-art performance on spatial deep learning tasks. They exhibit high data efficiency during training and significantly reduced inference time for interatomic potential calculations compared to classical approaches. Key to these models is the Clebsch-Gordon (CG) tensor product, a kernel that contracts two dense feature vectors with a highly structured sparse tensor to produce a dense output vector. The operation, which may be repeated millions of times for typical equivariant models, is a costly and inefficient bottleneck. We introduce a GPU sparse kernel generator for the CG tensor product that provides significant speedup over the best existing open and closed-source implementations. Our implementation achieves high performance by carefully managing GPU shared memory through static analysis at model compile-time, minimizing reads and writes to global memory. We break the tensor product into a series of kernels with operands that fit entirely into registers, enabling us to emit long arithmetic instruction streams that maximize instruction-level parallelism. By fusing the CG tensor product with a subsequent graph convolution, we reduce both intermediate storage and global memory traffic over naive approaches that duplicate input data. We also provide optimized kernels for the gradient of the CG tensor product and a novel identity for the higher partial derivatives required to predict interatomic forces. Our fused kernels offer up to 4.5x speedup for the forward pass and 3x for the backward pass over NVIDIA cuEquivariance, as well as >10x speedup over the widely-used e3nn package. We offer up to 5.3x inference-time speedup for the MACE chemistry foundation model over the original unoptimized version.
Authors: Sajjad Ghiasvand, Mahnoosh Alizadeh, Ramtin Pedarsani
Abstract: While parameter-efficient fine-tuning (PEFT) techniques like Low-Rank Adaptation (LoRA) offer computationally efficient adaptations of Large Language Models (LLMs), their practical deployment often assumes centralized data and training environments. However, real-world scenarios frequently involve distributed, privacy-sensitive datasets that require decentralized solutions. Federated learning (FL) addresses data privacy by coordinating model updates across clients, but it is typically based on centralized aggregation through a parameter server, which can introduce bottlenecks and communication constraints. Decentralized learning, in contrast, eliminates this dependency by enabling direct collaboration between clients, improving scalability and efficiency in distributed environments. Despite its advantages, decentralized LLM fine-tuning remains underexplored. In this work, we propose Dec-LoRA, a decentralized fine-tuning algorithm for LLMs based on LoRA. Through extensive experiments on BERT and LLaMA-2 models, we demonstrate that Dec-LoRA achieves performance comparable to centralized LoRA under various conditions, including data heterogeneity and quantization constraints. Additionally, we provide a rigorous theoretical guarantee proving the convergence of our algorithm to a stationary point for non-convex and smooth loss functions. These findings highlight the potential of Dec-LoRA for scalable LLM fine-tuning in decentralized environments.
Authors: Lucas Lange, Ole Borchardt, Erhard Rahm
Abstract: With growing concerns about user data collection, individualized privacy has emerged as a promising solution to balance protection and utility by accounting for diverse user privacy preferences. Instead of enforcing a uniform level of anonymization for all users, this approach allows individuals to choose privacy settings that align with their comfort levels. Building on this idea, we propose an adapted method for enabling Individualized Differential Privacy (IDP) in Federated Learning (FL) by handling clients according to their personal privacy preferences. By extending the SAMPLE algorithm from centralized settings to FL, we calculate client-specific sampling rates based on their heterogeneous privacy budgets and integrate them into a modified IDP-FedAvg algorithm. We test this method under realistic privacy distributions and multiple datasets. The experimental results demonstrate that our approach achieves clear improvements over uniform DP baselines, reducing the trade-off between privacy and utility. Compared to the alternative SCALE method in related work, which assigns differing noise scales to clients, our method performs notably better. However, challenges remain for complex tasks with non-i.i.d. data, primarily stemming from the constraints of the decentralized setting.
Authors: Giulio Masinelli, Chang Rajani, Patrik Hoffmann, Kilian Wasmer, David Atienza
Abstract: Ensuring consistent processing quality is challenging in laser processes due to varying material properties and surface conditions. Although some approaches have shown promise in solving this problem via automation, they often rely on predetermined targets or are limited to simulated environments. To address these shortcomings, we propose a novel real-time reinforcement learning approach for laser process control, implemented on a Field Programmable Gate Array to achieve real-time execution. Our experimental results from laser welding tests on stainless steel samples with a range of surface roughnesses validated the method's ability to adapt autonomously, without relying on reward engineering or prior setup information. Specifically, the algorithm learned the correct power profile for each unique surface characteristic, demonstrating significant improvements over hand-engineered optimal constant power strategies -- up to 23% better performance on rougher surfaces and 7% on mixed surfaces. This approach represents a significant advancement in automating and optimizing laser processes, with potential applications across multiple industries.
Authors: Ryotaro Kawata, Kazusato Oko, Atsushi Nitanda, Taiji Suzuki
Abstract: We introduce a novel alignment method for diffusion models from distribution optimization perspectives while providing rigorous convergence guarantees. We first formulate the problem as a generic regularized loss minimization over probability distributions and directly optimize the distribution using the Dual Averaging method. Next, we enable sampling from the learned distribution by approximating its score function via Doob's $h$-transform technique. The proposed framework is supported by rigorous convergence guarantees and an end-to-end bound on the sampling error, which imply that when the original distribution's score is known accurately, the complexity of sampling from shifted distributions is independent of isoperimetric conditions. This framework is broadly applicable to general distribution optimization problems, including alignment tasks in Reinforcement Learning with Human Feedback (RLHF), Direct Preference Optimization (DPO), and Kahneman-Tversky Optimization (KTO). We empirically validate its performance on synthetic and image datasets using the DPO objective.
Authors: Jaeyeon Kim, Kulin Shah, Vasilis Kontonis, Sham Kakade, Sitan Chen
Abstract: In recent years, masked diffusion models (MDMs) have emerged as a promising alternative approach for generative modeling over discrete domains. Compared to autoregressive models (ARMs), MDMs trade off complexity at training time with flexibility at inference time. At training time, they must learn to solve an exponentially large number of infilling problems, but at inference time, they can decode tokens in essentially arbitrary order. In this work, we closely examine these two competing effects. On the training front, we theoretically and empirically demonstrate that MDMs indeed train on computationally intractable subproblems compared to their autoregressive counterparts. On the inference front, we show that a suitable strategy for adaptively choosing the token decoding order significantly enhances the capabilities of MDMs, allowing them to sidestep hard subproblems. On logic puzzles like Sudoku, we show that adaptive inference can boost solving accuracy in pretrained MDMs from $<7$% to $\approx 90$%, even outperforming ARMs with $7\times$ as many parameters and that were explicitly trained via teacher forcing to learn the right order of decoding.
Authors: Buqing Nie, Yangqing Fu, Yue Gao
Abstract: Reinforcement Learning (RL) has achieved remarkable success in various continuous control tasks, such as robot manipulation and locomotion. Different to mainstream RL which makes decisions at individual steps, recent studies have incorporated action repetition into RL, achieving enhanced action persistence with improved sample efficiency and superior performance. However, existing methods treat all action dimensions as a whole during repetition, ignoring variations among them. This constraint leads to inflexibility in decisions, which reduces policy agility with inferior effectiveness. In this work, we propose a novel repetition framework called SDAR, which implements Spatially Decoupled Action Repetition through performing closed-loop act-or-repeat selection for each action dimension individually. SDAR achieves more flexible repetition strategies, leading to an improved balance between action persistence and diversity. Compared to existing repetition frameworks, SDAR is more sample efficient with higher policy performance and reduced action fluctuation. Experiments are conducted on various continuous control scenarios, demonstrating the effectiveness of spatially decoupled repetition design proposed in this work.
Authors: Hoony Kang, Wolfgang Losert
Abstract: The brain rapidly adapts to new contexts and learns from limited data, a coveted characteristic that artificial intelligence (AI) algorithms struggle to mimic. Inspired by the mechanical oscillatory rhythms of neural cells, we developed a learning paradigm utilizing link strength oscillations, where learning is associated with the coordination of these oscillations. Link oscillations can rapidly change coordination, allowing the network to sense and adapt to subtle contextual changes without supervision. The network becomes a generalist AI architecture, capable of predicting dynamics of multiple contexts including unseen ones. These results make our paradigm a powerful starting point for novel models of cognition. Because our paradigm is agnostic to specifics of the neural network, our study opens doors for introducing rapid adaptive learning into leading AI models.
Authors: Xiangfei Qiu, Hanyin Cheng, Xingjian Wu, Jilin Hu, Chenjuan Guo, Bin Yang
Abstract: Multivariate Time Series Forecasting (MTSF) plays a crucial role across diverse fields, ranging from economic, energy, to traffic. In recent years, deep learning has demonstrated outstanding performance in MTSF tasks. In MTSF, modeling the correlations among different channels is critical, as leveraging information from other related channels can significantly improve the prediction accuracy of a specific channel. This study systematically reviews the channel modeling strategies for time series and proposes a taxonomy organized into three hierarchical levels: the strategy perspective, the mechanism perspective, and the characteristic perspective. On this basis, we provide a structured analysis of these methods and conduct an in-depth examination of the advantages and limitations of different channel strategies. Finally, we summarize and discuss some future research directions to provide useful research guidance. Moreover, we maintain an up-to-date Github repository (https://github.com/decisionintelligence/CS4TS) which includes all the papers discussed in the survey.
Authors: Keir Adams, Connor W. Coley
Abstract: Training machine learning models to predict properties of molecular conformer ensembles is an increasingly popular strategy to accelerate the conformational analysis of drug-like small molecules, reactive organic substrates, and homogeneous catalysts. For high-throughput analyses especially, trained surrogate models can help circumvent traditional approaches to conformational analysis that rely on expensive conformer searches and geometry optimizations. Here, we question how the performance of surrogate models for predicting 3D conformer-dependent properties (of a single, active conformer) is affected by the quality of the 3D conformers used as their input. How well do lower-quality conformers inform the prediction of properties of higher-quality conformers? Does the fidelity of geometry optimization matter when encoding random conformers? For models that encode sets of conformers, how does the presence of the active conformer that induces the target property affect model accuracy? How do predictions from a surrogate model compare to estimating the properties from cheap ensembles themselves? We explore these questions in the context of predicting Sterimol parameters of conformer ensembles optimized with density functional theory. Although answers will be case-specific, our analyses provide a valuable perspective on 3D representation learning models and raise practical considerations regarding when conformer quality matters.
Authors: Eldad Haber, Shadab Ahamed, Md. Shahriar Rahim Siddiqui, Niloufar Zakariaei, Moshe Eliasof
Abstract: Generative models for image generation are now commonly used for a wide variety of applications, ranging from guided image generation for entertainment to solving inverse problems. Nonetheless, training a generator is a non-trivial feat that requires fine-tuning and can lead to so-called hallucinations, that is, the generation of images that are unrealistic. In this work, we explore image generation using flow matching. We explain and demonstrate why flow matching can generate hallucinations, and propose an iterative process to improve the generation process. Our iterative process can be integrated into virtually $\textit{any}$ generative modeling technique, thereby enhancing the performance and robustness of image synthesis systems.
Authors: Yancheng Zhang, Jiaqi Xue, Mengxin Zheng, Mimi Xie, Mingzhe Zhang, Lei Jiang, Qian Lou
Abstract: Private Transformer inference using cryptographic protocols offers promising solutions for privacy-preserving machine learning; however, it still faces significant runtime overhead (efficiency issues) and challenges in handling long-token inputs (scalability issues). We observe that the Transformer's operational complexity scales quadratically with the number of input tokens, making it essential to reduce the input token length. Notably, each token varies in importance, and many inputs contain redundant tokens. Additionally, prior private inference methods that rely on high-degree polynomial approximations for non-linear activations are computationally expensive. Therefore, reducing the polynomial degree for less important tokens can significantly accelerate private inference. Building on these observations, we propose \textit{CipherPrune}, an efficient and scalable private inference framework that includes a secure encrypted token pruning protocol, a polynomial reduction protocol, and corresponding Transformer network optimizations. At the protocol level, encrypted token pruning adaptively removes unimportant tokens from encrypted inputs in a progressive, layer-wise manner. Additionally, encrypted polynomial reduction assigns lower-degree polynomials to less important tokens after pruning, enhancing efficiency without decryption. At the network level, we introduce protocol-aware network optimization via a gradient-based search to maximize pruning thresholds and polynomial reduction conditions while maintaining the desired accuracy. Our experiments demonstrate that CipherPrune reduces the execution overhead of private Transformer inference by approximately $6.1\times$ for 128-token inputs and $10.6\times$ for 512-token inputs, compared to previous methods, with only a marginal drop in accuracy. The code is publicly available at https://github.com/UCF-Lou-Lab-PET/cipher-prune-inference.
URLs: https://github.com/UCF-Lou-Lab-PET/cipher-prune-inference.
Authors: Jacob Fein-Ashley
Abstract: Conventional self-attention mechanisms incur quadratic complexity, limiting their scalability on long sequences. We introduce \textbf{FFTNet}, an adaptive spectral filtering framework that leverages the Fast Fourier Transform (FFT) to achieve global token mixing in $\mathcal{O}(n\log n)$ time. By transforming inputs into the frequency domain, FFTNet exploits the orthogonality and energy preservation guaranteed by Parseval's theorem to capture long-range dependencies efficiently. Our main theoretical contributions are 1) an adaptive spectral filter, 2) combining local windowing with a global FFT branch, and 3) rich nonlinearity introduction in both the frequency and token domains. Experiments on the Long Range Arena and ImageNet benchmarks validate our theoretical insights and demonstrate superior performance over fixed Fourier and standard attention models.
Authors: Shashank Gupta, Chaitanya Ahuja, Tsung-Yu Lin, Sreya Dutta Roy, Harrie Oosterhuis, Maarten de Rijke, Satya Narayan Shukla
Abstract: Reinforcement learning (RL)-based fine-tuning has emerged as a powerful approach for aligning diffusion models with black-box objectives. Proximal policy optimization (PPO) is the most popular choice of method for policy optimization. While effective in terms of performance, PPO is highly sensitive to hyper-parameters and involves substantial computational overhead. REINFORCE, on the other hand, mitigates some computational complexities such as high memory overhead and sensitive hyper-parameter tuning, but has suboptimal performance due to high-variance and sample inefficiency. While the variance of the REINFORCE can be reduced by sampling multiple actions per input prompt and using a baseline correction term, it still suffers from sample inefficiency. To address these challenges, we systematically analyze the efficiency-effectiveness trade-off between REINFORCE and PPO, and propose leave-one-out PPO (LOOP), a novel RL for diffusion fine-tuning method. LOOP combines variance reduction techniques from REINFORCE, such as sampling multiple actions per input prompt and a baseline correction term, with the robustness and sample efficiency of PPO via clipping and importance sampling. Our results demonstrate that LOOP effectively improves diffusion models on various black-box objectives, and achieves a better balance between computational efficiency and performance.
Authors: Bo Yang
Abstract: Large language models (LLMs) inadvertently memorize sensitive data from their massive pretraining corpora \cite{jang2022knowledge}. In this work, we propose CE-U (Cross Entropy Unlearning), a novel loss function designed specifically for unlearning tasks. CE-U addresses fundamental limitations of gradient ascent approaches which suffer from instability due to vanishing gradients when model confidence is high and gradient exploding when confidence is low. We also unify standard cross entropy supervision and cross entropy unlearning into a single framework. Notably, on the TOFU benchmark for unlearning \cite{maini2024tofu}, CE-U achieves state-of-the-art results on LLaMA2-7B with 1\% and 5\% forgetting, even without the use of any extra reference model or additional positive samples. Our theoretical analysis further reveals that the gradient instability issues also exist in popular reinforcement learning algorithms like DPO \cite{rafailov2023direct} and GRPO\cite{Shao2024DeepSeekMath}, as they include a gradient ascent component. This suggests that applying CE-U principles to reinforcement learning could be a promising direction for improving stability and convergence.
Authors: Dayal Singh Kalra, John Kirchenbauer, Maissam Barkeshli, Tom Goldstein
Abstract: Adam is the go-to optimizer for training modern machine learning models, but it requires additional memory to maintain the moving averages of the gradients and their squares. While various low-memory optimizers have been proposed that sometimes match the performance of Adam, their lack of reliability has left Adam as the default choice. In this work, we apply a simple layer-wise Signal-to-Noise Ratio (SNR) analysis to quantify when second-moment tensors can be effectively replaced by their means across different dimensions. Our SNR analysis reveals how architecture, training hyperparameters, and dataset properties impact compressibility along Adam's trajectory, naturally leading to $\textit{SlimAdam}$, a memory-efficient Adam variant. $\textit{SlimAdam}$ compresses the second moments along dimensions with high SNR when feasible, and leaves when compression would be detrimental. Through experiments across a diverse set of architectures and training scenarios, we show that $\textit{SlimAdam}$ matches Adam's performance and stability while saving up to $98\%$ of total second moments. Code for $\textit{SlimAdam}$ is available at https://github.com/dayal-kalra/low-memory-adam.
Authors: Leonid Berlyand, Theo Bourdais, Houman Owhadi, Yitzchak Shmalo
Abstract: Deep neural networks (DNNs) have brought significant advancements in various applications in recent years, such as image recognition, speech recognition, and natural language processing. In particular, Vision Transformers (ViTs) have emerged as a powerful class of models in the field of deep learning for image classification. In this work, we propose a novel Random Matrix Theory (RMT)-based method for pruning pre-trained DNNs, based on the sparsification of weights and singular vectors, and apply it to ViTs. RMT provides a robust framework to analyze the statistical properties of large matrices, which has been shown to be crucial for understanding and optimizing the performance of DNNs. We demonstrate that our RMT-based pruning can be used to reduce the number of parameters of ViT models (trained on ImageNet) by 30-50\% with less than 1\% loss in accuracy. To our knowledge, this represents the state-of-the-art in pruning for these ViT models. Furthermore, we provide a rigorous mathematical underpinning of the above numerical studies, namely we proved a theorem for fully connected DNNs, and other more general DNN structures, describing how the randomness in the weight matrices of a DNN decreases as the weights approach a local or global minimum (during training). We verify this theorem through numerical experiments on fully connected DNNs, providing empirical support for our theoretical findings. Moreover, we prove a theorem that describes how DNN loss decreases as we remove randomness in the weight layers, and show a monotone dependence of the decrease in loss with the amount of randomness that we remove. Our results also provide significant RMT-based insights into the role of regularization during training and pruning.
Authors: Haoxiang You, Lekan Molu, Ian Abraham
Abstract: The Bellman equation and its continuous-time counterpart, the Hamilton-Jacobi-Bellman (HJB) equation, serve as necessary conditions for optimality in reinforcement learning and optimal control. While the value function is known to be the unique solution to the Bellman equation in tabular settings, we demonstrate that this uniqueness fails to hold in continuous state spaces. Specifically, for linear dynamical systems, we prove the Bellman equation admits at least $\binom{2n}{n}$ solutions, where $n$ is the state dimension. Crucially, only one of these solutions yields both an optimal policy and a stable closed-loop system. We then demonstrate a common failure mode in value-based methods: convergence to unstable solutions due to the exponential imbalance between admissible and inadmissible solutions. Finally, we introduce a positive-definite neural architecture that guarantees convergence to the stable solution by construction to address this issue.
Authors: Qiyi Wang, Yinning Shao, Yunlong Ma, Min Liu
Abstract: Graph neural architecture search (GraphNAS) has demonstrated advantages in mitigating performance degradation of graph neural networks (GNNs) due to distribution shifts. Recent approaches introduce weight sharing across tailored architectures, generating unique GNN architectures for each graph end-to-end. However, existing GraphNAS methods do not account for distribution patterns across different graphs and heavily rely on extensive training data. With sparse or single training graphs, these methods struggle to discover optimal mappings between graphs and architectures, failing to generalize to out-of-distribution (OOD) data. In this paper, we propose node-specific graph neural architecture search(NodeNAS), which aims to tailor distinct aggregation methods for different nodes through disentangling node topology and graph distribution with limited datasets. We further propose adaptive aggregation attention based Multi-dim NodeNAS method(MNNAS), which learns an node-specific architecture customizer with good generalizability. Specifically, we extend the vertical depth of the search space, supporting simultaneous node-specific architecture customization across multiple dimensions. Moreover, we model the power-law distribution of node degrees under varying assortativity, encoding structure invariant information to guide architecture customization across each dimension. Extensive experiments across supervised and unsupervised tasks demonstrate that MNNAS surpasses state-of-the-art algorithms and achieves excellent OOD generalization.
Authors: Yujiao Yang, Jing Lian, Linhui Li
Abstract: We propose Union-of-Experts (UoE), which decomposes transformer into an equitant group of experts, and then implement selective routing on input data and experts. Our approach advances MoE design with four key innovations: (1) We conducted equitant expert decomposition on both MLP blocks and attention blocks based on matrix partition in tensor parallelism. (2) We developed two routing paradigms: patch-wise data selection and expert selection, to apply routing across different levels. (3) We design the architecture of UoE model, including Selective Multi-Head Attention (SMHA) and Union-of-MLP-Experts (UoME). (4) We develop parallel implementation of UoE's routing and computation operation, and optimize efficiency based on the hardware processing analysis. The experiments demonstrate that the UoE model surpass Full Attention, state-of-art MoEs and efficient transformers (including the model architecture of recently proposed DeepSeek-V3) in several tasks across image and natural language domains. In language modeling tasks, we achieve an average reduction of 2.38 in perplexity compared to the best-performed MoE method with an average of 76% FLOPs. In Long Range Arena benchmark, we recorded an average score that is at least 0.68% higher than all comparison models including Full Attention, MoEs, and transformer variants, with only 50% FLOPs of the best MoE method. In image classification, our model yielded an average accuracy improvement of 1.75% than the best model while maintaining comparable FLOPs. The source codes are available at https://github.com/YujiaoYang-work/UoE.
Authors: Alicia Russell-Gilbert, Sudip Mittal, Shahram Rahimi, Maria Seale, Joseph Jabour, Thomas Arnold, Joshua Church
Abstract: Anomaly detection in complex industrial environments poses unique challenges, particularly in contexts characterized by data sparsity and evolving operational conditions. Predictive maintenance (PdM) in such settings demands methodologies that are adaptive, transferable, and capable of integrating domain-specific knowledge. In this paper, we present RAAD-LLM, a novel framework for adaptive anomaly detection, leveraging large language models (LLMs) integrated with Retrieval-Augmented Generation (RAG). This approach addresses the aforementioned PdM challenges. By effectively utilizing domain-specific knowledge, RAAD-LLM enhances the detection of anomalies in time series data without requiring fine-tuning on specific datasets. The framework's adaptability mechanism enables it to adjust its understanding of normal operating conditions dynamically, thus increasing detection accuracy. We validate this methodology through a real-world application for a plastics manufacturing plant and the Skoltech Anomaly Benchmark (SKAB). Results show significant improvements over our previous model with an accuracy increase from 70.7% to 89.1% on the real-world dataset. By allowing for the enriching of input series data with semantics, RAAD-LLM incorporates multimodal capabilities that facilitate more collaborative decision-making between the model and plant operators. Overall, our findings support RAAD-LLM's ability to revolutionize anomaly detection methodologies in PdM, potentially leading to a paradigm shift in how anomaly detection is implemented across various industries.
Authors: Vaibhav Singh, Paul Janson, Paria Mehrbod, Adam Ibrahim, Irina Rish, Eugene Belilovsky, Benjamin Th\'erien
Abstract: The ever-growing availability of unlabeled data presents both opportunities and challenges for training artificial intelligence systems. While self-supervised learning (SSL) has emerged as a powerful paradigm for extracting meaningful representations from vast amounts of unlabeled data, existing methods still struggle to adapt to the non-stationary, non-IID nature of real-world data streams without forgetting previously learned knowledge. Recent works have adopted a repeated cosine annealing schedule for large-scale continual pre-training; however, these schedules (1) inherently cause forgetting during the re-warming phase and (2) have not been systematically compared to existing continual SSL methods. In this work, we systematically compare the widely used cosine schedule with the recently proposed infinite learning rate schedule and empirically find the latter to be a more effective alternative. Our extensive empirical evaluation across diverse image and language datasets demonstrates that the infinite learning rate schedule consistently enhances continual pre-training performance compared to a repeated cosine decay without being restricted to a fixed iteration budget. For instance, in a small-scale MAE pre-training setup, it outperforms several strong baselines from the literature. We then scale up our experiments to larger MAE pre-training and autoregressive language model pre-training. Our results show that the infinite learning rate schedule remains effective at scale, surpassing repeated cosine decay for both MAE pre-training and zero-shot LM benchmarks.
Authors: Alexander Kolpakov, Igor Rivin
Abstract: DiRe - JAX is a new dimensionality reduction toolkit designed to address some of the challenges faced by traditional methods like UMAP and tSNE such as loss of global structure and computational efficiency. Built on the JAX framework, DiRe leverages modern hardware acceleration to provide an efficient, scalable, and interpretable solution for visualizing complex data structures, and for quantitative analysis of lower-dimensional embeddings. The toolkit shows considerable promise in preserving both local and global structures within the data as compared to state-of-the-art UMAP and tSNE implementations. This makes it suitable for a wide range of applications in machine learning, bio-informatics, and data science.
Authors: Dong Tian, Ge Li, Hongyi Zhou, Onur Celik, Gerhard Neumann
Abstract: Soft Actor-Critic (SAC) critically depends on its critic network, which typically evaluates a single state-action pair to guide policy updates. Using N-step returns is a common practice to reduce the bias in the target values of the critic. However, using N-step returns can again introduce high variance and necessitates importance sampling, often destabilizing training. Recent algorithms have also explored action chunking-such as direct action repetition and movement primitives-to enhance exploration. In this paper, we propose a Transformer-based Critic Network for SAC that integrates the N-returns framework in a stable and efficient manner. Unlike approaches that perform chunking in the actor network, we feed chunked actions into the critic network to explore potential performance gains. Our architecture leverages the Transformer's ability to process sequential information, facilitating more robust value estimation. Empirical results show that this method not only achieves efficient, stable training but also excels in sparse reward/multi-phase environments-traditionally a challenge for step-based methods. These findings underscore the promise of combining Transformer-based critics with N-returns to advance reinforcement learning performance
Authors: Shaojun Wu, Shan Jin, Dingding Wen, Donghong Han, Xiaoting Wang
Abstract: Quantum reinforcement learning (QRL) is a promising paradigm for near-term quantum devices. While existing QRL methods have shown success in discrete action spaces, extending these techniques to continuous domains is challenging due to the curse of dimensionality introduced by discretization. To overcome this limitation, we introduce a quantum Deep Deterministic Policy Gradient (DDPG) algorithm that efficiently addresses both classical and quantum sequential decision problems in continuous action spaces. Moreover, our approach facilitates single-shot quantum state generation: a one-time optimization produces a model that outputs the control sequence required to drive a fixed initial state to any desired target state. In contrast, conventional quantum control methods demand separate optimization for each target state. We demonstrate the effectiveness of our method through simulations and discuss its potential applications in quantum control.
Authors: Sara Fish, Paul G\"olz, David C. Parkes, Ariel D. Procaccia, Gili Rusak, Itai Shapira, Manuel W\"uthrich
Abstract: The mathematical study of voting, social choice theory, has traditionally only been applicable to choices among a few predetermined alternatives, but not to open-ended decisions such as collectively selecting a textual statement. We introduce generative social choice, a design methodology for open-ended democratic processes that combines the rigor of social choice theory with the capability of large language models to generate text and extrapolate preferences. Our framework divides the design of AI-augmented democratic processes into two components: first, proving that the process satisfies representation guarantees when given access to oracle queries; second, empirically validating that these queries can be approximately implemented using a large language model. We apply this framework to the problem of summarizing free-form opinions into a proportionally representative slate of opinion statements; specifically, we develop a democratic process with representation guarantees and use this process to portray the opinions of participants in a survey about abortion policy. In a trial with 100 representative US residents, we find that 84 out of 100 participants feel "excellently" or "exceptionally" represented by the slate of five statements we extracted.
Authors: Andrea Angiuli, Jean-Pierre Fouque, Ruimeng Hu, Alan Raydan
Abstract: We present the development and analysis of a reinforcement learning (RL) algorithm designed to solve continuous-space mean field game (MFG) and mean field control (MFC) problems in a unified manner. The proposed approach pairs the actor-critic (AC) paradigm with a representation of the mean field distribution via a parameterized score function, which can be efficiently updated in an online fashion, and uses Langevin dynamics to obtain samples from the resulting distribution. The AC agent and the score function are updated iteratively to converge, either to the MFG equilibrium or the MFC optimum for a given mean field problem, depending on the choice of learning rates. A straightforward modification of the algorithm allows us to solve mixed mean field control games (MFCGs). The performance of our algorithm is evaluated using linear-quadratic benchmarks in the asymptotic infinite horizon framework.
Authors: Kyriakos Flouris, Ender Konukoglu
Abstract: Manifold learning flows are a class of generative modelling techniques that assume a low-dimensional manifold description of the data. The embedding of such a manifold into the high-dimensional space of the data is achieved via learnable invertible transformations. Therefore, once the manifold is properly aligned via a reconstruction loss, the probability density is tractable on the manifold and maximum likelihood can be used to optimize the network parameters. Naturally, the lower-dimensional representation of the data requires an injective-mapping. Recent approaches were able to enforce that the density aligns with the modelled manifold, while efficiently calculating the density volume-change term when embedding to the higher-dimensional space. However, unless the injective-mapping is analytically predefined, the learned manifold is not necessarily an efficient representation of the data. Namely, the latent dimensions of such models frequently learn an entangled intrinsic basis, with degenerate information being stored in each dimension. Alternatively, if a locally orthogonal and/or sparse basis is to be learned, here coined canonical intrinsic basis, it can serve in learning a more compact latent space representation. Toward this end, we propose a canonical manifold learning flow method, where a novel optimization objective enforces the transformation matrix to have few prominent and non-degenerate basis functions. We demonstrate that by minimizing the off-diagonal manifold metric elements $\ell_1$-norm, we can achieve such a basis, which is simultaneously sparse and/or orthogonal. Canonical manifold flow yields a more efficient use of the latent space, automatically generating fewer prominent and distinct dimensions to represent data, and a better approximation of target distributions than other manifold flow methods in most experiments we conducted, resulting in lower FID scores.
Authors: Jessica Ojo, Odunayo Ogundepo, Akintunde Oladipo, Kelechi Ogueji, Jimmy Lin, Pontus Stenetorp, David Ifeoluwa Adelani
Abstract: Large-scale multilingual evaluations, such as MEGA, often include only a handful of African languages due to the scarcity of high-quality evaluation data and the limited discoverability of existing African datasets. This lack of representation hinders comprehensive LLM evaluation across a diverse range of languages and tasks. To address these challenges, we introduce AfroBench -- a multi-task benchmark for evaluating the performance of LLMs across 64 African languages, 15 tasks and 22 datasets. AfroBench consists of nine natural language understanding datasets, six text generation datasets, six knowledge and question answering tasks, and one mathematical reasoning task. We present results comparing the performance of prompting LLMs to fine-tuned baselines based on BERT and T5-style models. Our results suggest large gaps in performance between high-resource languages, such as English, and African languages across most tasks; but performance also varies based on the availability of monolingual data resources. Our findings confirm that performance on African languages continues to remain a hurdle for current LLMs, underscoring the need for additional efforts to close this gap. https://mcgill-nlp.github.io/AfroBench/
Authors: Baris Ata, Ebru Kasikaralar
Abstract: We consider a multi-class queueing model of a telephone call center, in which a system manager dynamically allocates available servers to customer calls. Calls can terminate through either service completion or customer abandonment, and the manager strives to minimize the expected total of holding costs plus abandonment costs over a finite horizon. Focusing on the Halfin-Whitt heavy traffic regime, we derive an approximating diffusion control problem, and building on earlier work by Beck et al. (2021), develop a simulation-based computational method for solution of such problems, one that relies heavily on deep neural network technology. Using this computational method, we propose a policy for the original (pre-limit) call center scheduling problem. Finally, the performance of this policy is assessed using test problems based on publicly available call center data. For the test problems considered so far, our policy does as well as or better than the best benchmark we could find. Moreover, our method is computationally feasible at least up to dimension 500, that is, for call centers with 500 or more distinct customer classes.
Authors: Marco Fanizza, Niklas Galke, Josep Lumbreras, Cambyse Rouz\'e, Andreas Winter
Abstract: Matrix product operators allow efficient descriptions (or realizations) of states on a 1D lattice. We consider the task of learning a realization of minimal dimension from copies of an unknown state, such that the resulting operator is close to the density matrix in trace norm. For finitely correlated translation-invariant states on an infinite chain, a realization of minimal dimension can be exactly reconstructed via linear algebra operations from the marginals of a size depending on the representation dimension. We establish a bound on the trace norm error for an algorithm that estimates a candidate realization from estimates of these marginals and outputs a matrix product operator, estimating the state of a chain of arbitrary length $t$. This bound allows us to establish an $O(t^2)$ upper bound on the sample complexity of the learning task, with an explicit dependence on the site dimension, realization dimension and spectral properties of a certain map constructed from the state. A refined error bound can be proven for $C^*$-finitely correlated states, which have an operational interpretation in terms of sequential quantum channels applied to the memory system. We can also obtain an analogous error bound for a class of matrix product density operators on a finite chain reconstructible by local marginals. In this case, a linear number of marginals must be estimated, obtaining a sample complexity of $\tilde{O}(t^3)$. The learning algorithm also works for states that are sufficiently close to a finitely correlated state, with the potential of providing competitive algorithms for other interesting families of states.
Authors: Pablo Lemos, Sammy Sharief, Nikolay Malkin, Salma Salhi, Conner Stone, Laurence Perreault-Levasseur, Yashar Hezaveh
Abstract: We propose a likelihood-free method for comparing two distributions given samples from each, with the goal of assessing the quality of generative models. The proposed approach, PQMass, provides a statistically rigorous method for assessing the performance of a single generative model or the comparison of multiple competing models. PQMass divides the sample space into non-overlapping regions and applies chi-squared tests to the number of data samples that fall within each region, giving a p-value that measures the probability that the bin counts derived from two sets of samples are drawn from the same multinomial distribution. PQMass does not depend on assumptions regarding the density of the true distribution, nor does it rely on training or fitting any auxiliary models. We evaluate PQMass on data of various modalities and dimensions, demonstrating its effectiveness in assessing the quality, novelty, and diversity of generated samples. We further show that PQMass scales well to moderately high-dimensional data and thus obviates the need for feature extraction in practical applications.
Authors: Lele Cao, Valentin Buchner, Zineb Senane, Fangkai Yang
Abstract: Multimodal Large Language Models (MLLMs) are typically assessed using expensive annotated multimodal benchmarks, which often lag behind the rapidly evolving demands of MLLM evaluation. This paper outlines and validates GenCeption, a novel, annotation-free evaluation method that requires only unimodal data to measure inter-modality semantic coherence and inversely assesses MLLMs' tendency to hallucinate. This approach eliminates the need for costly data annotation, minimizes the risk of training data contamination, is expected to result in slower benchmark saturation, and avoids the illusion of emerging abilities. Inspired by the DrawCeption game, GenCeption begins with a non-textual sample and proceeds through iterative description and generation steps. The semantic drift across iterations is quantified using the GC@T metric. While GenCeption is principally applicable to MLLMs across various modalities, this paper focuses on its implementation and validation for Vision LLMs (VLLMs). Based on the GenCeption method, we establish the MMECeption benchmark for evaluating VLLMs, and compare the performance of several popular VLLMs and human annotators. Our empirical results validate GenCeption's effectiveness, demonstrating strong correlations with established VLLM benchmarks. VLLMs still significantly lag behind human performance and struggle especially with text-intensive tasks.
Authors: Paul Gavrikov, Jovita Lukasik, Steffen Jung, Robert Geirhos, M. Jehanzeb Mirza, Margret Keuper, Janis Keuper
Abstract: Unlike traditional vision-only models, vision language models (VLMs) offer an intuitive way to access visual content through language prompting by combining a large language model (LLM) with a vision encoder. However, both the LLM and the vision encoder come with their own set of biases, cue preferences, and shortcuts, which have been rigorously studied in uni-modal models. A timely question is how such (potentially misaligned) biases and cue preferences behave under multi-modal fusion in VLMs. As a first step towards a better understanding, we investigate a particularly well-studied vision-only bias - the texture vs. shape bias and the dominance of local over global information. As expected, we find that VLMs inherit this bias to some extent from their vision encoders. Surprisingly, the multi-modality alone proves to have important effects on the model behavior, i.e., the joint training and the language querying change the way visual cues are processed. While this direct impact of language-informed training on a model's visual perception is intriguing, it raises further questions on our ability to actively steer a model's output so that its prediction is based on particular visual cues of the user's choice. Interestingly, VLMs have an inherent tendency to recognize objects based on shape information, which is different from what a plain vision encoder would do. Further active steering towards shape-based classifications through language prompts is however limited. In contrast, active VLM steering towards texture-based decisions through simple natural language prompts is often more successful. URL: https://github.com/paulgavrikov/vlm_shapebias
Authors: Gilles Blanchard (LMO, DATASHAPE), Jean-Baptiste Fermanian (LMO), Hannah Marienwald (TUB)
Abstract: We endeavour to estimate numerous multi-dimensional means of various probability distributions on a common space based on independent samples. Our approach involves forming estimators through convex combinations of empirical means derived from these samples. We introduce two strategies to find appropriate data-dependent convex combination weights: a first one employing a testing procedure to identify neighbouring means with low variance, which results in a closed-form plug-in formula for the weights, and a second one determining weights via minimization of an upper confidence bound on the quadratic risk.Through theoretical analysis, we evaluate the improvement in quadratic risk offered by our methods compared to the empirical means. Our analysis focuses on a dimensional asymptotics perspective, showing that our methods asymptotically approach an oracle (minimax) improvement as the effective dimension of the data increases.We demonstrate the efficacy of our methods in estimating multiple kernel mean embeddings through experiments on both simulated and real-world datasets.
Authors: Younes Belkouchi, Jean-Christophe Pesquet, Audrey Repetti, Hugues Talbot
Abstract: This article introduces a novel approach to learning monotone neural networks through a newly defined penalization loss. The proposed method is particularly effective in solving classes of variational problems, specifically monotone inclusion problems, commonly encountered in image processing tasks. The Forward-Backward-Forward (FBF) algorithm is employed to address these problems, offering a solution even when the Lipschitz constant of the neural network is unknown. Notably, the FBF algorithm provides convergence guarantees under the condition that the learned operator is monotone. Building on plug-and-play methodologies, our objective is to apply these newly learned operators to solving non-linear inverse problems. To achieve this, we initially formulate the problem as a variational inclusion problem. Subsequently, we train a monotone neural network to approximate an operator that may not inherently be monotone. Leveraging the FBF algorithm, we then show simulation examples where the non-linear inverse problem is successfully solved.
Authors: Andi Zhang, Tim Z. Xiao, Weiyang Liu, Robert Bamler, Damon Wischik
Abstract: We revisit the likelihood ratio between a pretrained large language model (LLM) and its finetuned variant as a criterion for out-of-distribution (OOD) detection. The intuition behind such a criterion is that, the pretrained LLM has the prior knowledge about OOD data due to its large amount of training data, and once finetuned with the in-distribution data, the LLM has sufficient knowledge to distinguish their difference. Leveraging the power of LLMs, we show that, the likelihood ratio can serve as an effective OOD detection criterion. Moreover, we apply the proposed LLM-based likelihood ratio to detect OOD questions in question-answering (QA) systems, which can be used to improve the performance of specialized LLMs for general questions. Given that likelihood can be easily obtained by the loss functions within contemporary neural network frameworks, it is straightforward to implement this approach in practice. Since both the pretrained LLMs and its various finetuned models are widely available from online platforms such as Hugging Face, our proposed criterion can be effortlessly incorporated for OOD detection without the need for further training. We conduct comprehensive evaluation across on multiple settings, including far OOD, near OOD, spam detection, and QA scenarios, to demonstrate the effectiveness of the method. Code can be found at https://github.com/andiac/LLMOODratio
Authors: Abhinav Lalwani, Tasha Kim, Lovish Chopra, Christopher Hahn, Zhijing Jin, Mrinmaya Sachan
Abstract: Translating natural language into formal language such as First-Order Logic (FOL) is a foundational challenge in NLP with wide-ranging applications in automated reasoning, misinformation tracking, and knowledge validation. In this paper, we introduce Natural Language to First-Order Logic (NL2FOL), a framework to autoformalize natural language to FOL step by step using Large Language Models (LLMs). Our approach addresses key challenges in this translation process, including the integration of implicit background knowledge. By leveraging structured representations generated by NL2FOL, we use Satisfiability Modulo Theory (SMT) solvers to reason about the logical validity of natural language statements. We present logical fallacy detection as a case study to evaluate the efficacy of NL2FOL. Being neurosymbolic, our approach also provides interpretable insights into the reasoning process and demonstrates robustness without requiring model fine-tuning or labeled training data. Our framework achieves strong performance on multiple datasets. On the LOGIC dataset, NL2FOL achieves an F1-score of 78%, while generalizing effectively to the LOGICCLIMATE dataset with an F1-score of 80%.
Authors: Xinyi Shang, Peng Sun, Tao Lin
Abstract: Recent advancements in dataset distillation have demonstrated the significant benefits of employing soft labels generated by pre-trained teacher models. In this paper, we introduce a novel perspective by emphasizing the full utilization of labels. We first conduct a comprehensive comparison of various loss functions for soft label utilization in dataset distillation, revealing that the model trained on the synthetic dataset exhibits high sensitivity to the choice of loss function for soft label utilization. This finding highlights the necessity of a universal loss function for training models on synthetic datasets. Building on these insights, we introduce an extremely simple yet surprisingly effective plug-and-play approach, GIFT, which encompasses soft label refinement and a cosine similarity-based loss function to efficiently leverage full label information. Extensive experiments indicate that GIFT consistently enhances state-of-the-art dataset distillation methods across various dataset scales, without incurring additional computational costs. Importantly, GIFT significantly enhances cross-optimizer generalization, an area previously overlooked. For instance, on ImageNet-1K with IPC = 10, GIFT enhances the state-of-the-art method RDED by 30.8% in cross-optimizer generalization. Our code is available at https://github.com/LINs-lab/GIFT.
Authors: Aditya Ravuri, Neil D. Lawrence
Abstract: This paper shows that dimensionality reduction methods such as UMAP and t-SNE, can be approximately recast as MAP inference methods corresponding to a model introduced in ProbDR, that describes the graph Laplacian (an estimate of the data precision matrix) using a Wishart distribution, with a mean given by a non-linear covariance function evaluated on the latents. This interpretation offers deeper theoretical and semantic insights into such algorithms, by showing that variances corresponding to these covariances are low (potentially misspecified), and forging a connection to Gaussian process latent variable models by showing that well-known kernels can be used to describe covariances implied by graph Laplacians. We also introduce tools with which similar dimensionality reduction methods can be studied.
Authors: Arthur Jacot, Alexandre Kaiser
Abstract: We study Leaky ResNets, which interpolate between ResNets and Fully-Connected nets depending on an 'effective depth' hyper-parameter $\tilde{L}$. In the infinite depth limit, we study 'representation geodesics' $A_{p}$: continuous paths in representation space (similar to NeuralODEs) from input $p=0$ to output $p=1$ that minimize the parameter norm of the network. We give a Lagrangian and Hamiltonian reformulation, which highlight the importance of two terms: a kinetic energy which favors small layer derivatives $\partial_{p}A_{p}$ and a potential energy that favors low-dimensional representations, as measured by the 'Cost of Identity'. The balance between these two forces offers an intuitive understanding of feature learning in ResNets. We leverage this intuition to explain the emergence of a bottleneck structure, as observed in previous work: for large $\tilde{L}$ the potential energy dominates and leads to a separation of timescales, where the representation jumps rapidly from the high dimensional inputs to a low-dimensional representation, move slowly inside the space of low-dimensional representations, before jumping back to the potentially high-dimensional outputs. Inspired by this phenomenon, we train with an adaptive layer step-size to adapt to the separation of timescales.
Authors: Yixuan Mei, Yonghao Zhuang, Xupeng Miao, Juncheng Yang, Zhihao Jia, Rashmi Vinayak
Abstract: This paper introduces Helix, a distributed system for high-throughput, low-latency large language model (LLM) serving in heterogeneous GPU clusters. The key idea behind Helix is to formulate inference computation of LLMs over heterogeneous GPUs and network connections as a max-flow problem on directed, weighted graphs, whose nodes represent GPU instances and edges capture both GPU and network heterogeneity through their capacities. Helix then uses a mixed integer linear programming (MILP) algorithm to discover highly optimized strategies to serve LLMs on heterogeneous GPUs. This approach allows Helix to jointly optimize model placement and request scheduling, two highly entangled tasks in heterogeneous LLM serving. Our evaluation on several heterogeneous clusters ranging from 24 to 42 GPU nodes shows that Helix improves serving throughput by up to 3.3x and reduces prompting and decoding latency by up to 66% and 24%, respectively, compared to existing approaches. Helix is available at https://github.com/Thesys-lab/Helix-ASPLOS25.
Authors: Arya Akhavan, Karim Lounici, Massimiliano Pontil, Alexandre B. Tsybakov
Abstract: We study the contextual continuum bandits problem, where the learner sequentially receives a side information vector and has to choose an action in a convex set, minimizing a function associated with the context. The goal is to minimize all the underlying functions for the received contexts, leading to the contextual notion of regret, which is stronger than the standard static regret. Assuming that the objective functions are $\gamma$-H\"older with respect to the contexts, $0<\gamma\le 1,$ we demonstrate that any algorithm achieving a sub-linear static regret can be extended to achieve a sub-linear contextual regret. We prove a static-to-contextual regret conversion theorem that provides an upper bound for the contextual regret of the output algorithm as a function of the static regret of the input algorithm. We further study the implications of this general result for three fundamental cases of dependency of the objective function on the action variable: (a) Lipschitz bandits, (b) convex bandits, (c) strongly convex and smooth bandits. For Lipschitz bandits and $\gamma=1,$ combining our results with the lower bound of Slivkins (2014), we prove that the minimax optimal contextual regret for the noise-free adversarial setting is achieved. Then, we prove that in the presence of noise, the contextual regret rate as a function of the number of queries is the same for convex bandits as it is for strongly convex and smooth bandits. Lastly, we present a minimax lower bound, implying two key facts. First, obtaining a sub-linear contextual regret may be impossible over functions that are not continuous with respect to the context. Second, for convex bandits and strongly convex and smooth bandits, the algorithms that we propose achieve, up to a logarithmic factor, the minimax optimal rate of contextual regret as a function of the number of queries.
Authors: Chufan Gao, Jathurshan Pradeepkumar, Trisha Das, Shivashankar Thati, Jimeng Sun
Abstract: Background The cost of drug discovery and development is substantial, with clinical trial outcomes playing a critical role in regulatory approval and patient care. However, access to large-scale, high-quality clinical trial outcome data remains limited, hindering advancements in predictive modeling and evidence-based decision-making. Methods We present the Clinical Trial Outcome (CTO) benchmark, a fully reproducible, large-scale repository encompassing approximately 125,000 drug and biologics trials. CTO integrates large language model (LLM) interpretations of publications, trial phase progression tracking, sentiment analysis from news sources, stock price movements of trial sponsors, and additional trial-related metrics. Furthermore, we manually annotated a dataset of clinical trials conducted between 2020 and 2024 to enhance the quality and reliability of outcome labels. Results The trial outcome labels in the CTO benchmark agree strongly with expert annotations, achieving an F1 score of 94 for Phase 3 trials and 91 across all phases. Additionally, benchmarking standard machine learning models on our manually annotated dataset revealed distribution shifts in recent trials, underscoring the necessity of continuously updated labeling approaches. Conclusions By analyzing CTO's performance on recent clinical trials, we demonstrate the ongoing need for high-quality, up-to-date trial outcome labels. We publicly release the CTO knowledge base and annotated labels at https://chufangao.github.io/CTOD, with regular updates to support research on clinical trial outcomes and inform data-driven improvements in drug development.
Authors: Arthur Jacot, Seok Hoan Choi, Yuxiao Wen
Abstract: We show that deep neural networks (DNNs) can efficiently learn any composition of functions with bounded $F_{1}$-norm, which allows DNNs to break the curse of dimensionality in ways that shallow networks cannot. More specifically, we derive a generalization bound that combines a covering number argument for compositionality, and the $F_{1}$-norm (or the related Barron norm) for large width adaptivity. We show that the global minimizer of the regularized loss of DNNs can fit for example the composition of two functions $f^{*}=h\circ g$ from a small number of observations, assuming $g$ is smooth/regular and reduces the dimensionality (e.g. $g$ could be the quotient map of the symmetries of $f^{*}$), so that $h$ can be learned in spite of its low regularity. The measures of regularity we consider is the Sobolev norm with different levels of differentiability, which is well adapted to the $F_{1}$ norm. We compute scaling laws empirically and observe phase transitions depending on whether $g$ or $h$ is harder to learn, as predicted by our theory.
Authors: Oladipo A. Madamidola, Felix Ngobigha, Adnane Ez-zizi
Abstract: Machine learning has been successfully applied in developing malware detection systems, with a primary focus on accuracy, and increasing attention to reducing computational overhead and improving model interpretability. However, an important question remains underexplored: How well can machine learning-based models detect entirely new forms of malware not present in the training data? In this study, we present a machine learning-based system for detecting obfuscated malware that is not only highly accurate, lightweight and interpretable, but also capable of successfully adapting to new types of malware attacks. Our system is capable of detecting 15 malware subtypes despite being exclusively trained on one malware subtype, namely the Transponder from the Spyware family. This system was built after training 15 distinct random forest-based models, each on a different malware subtype from the CIC-MalMem-2022 dataset. These models were evaluated against the entire range of malware subtypes, including all unseen malware subtypes. To maintain the system's streamlined nature, training was confined to the top five most important features, which also enhanced interpretability. The Transponder-focused model exhibited high accuracy, exceeding 99.8%, with an average processing speed of 5.7 microseconds per file. We also illustrate how the Shapley additive explanations technique can facilitate the interpretation of the model predictions. Our research contributes to advancing malware detection methodologies, pioneering the feasibility of detecting obfuscated malware by exclusively training a model on a single or a few carefully selected malware subtypes and applying it to detect unseen subtypes.
Authors: Joshua Zhi En Tan, JunJie Wee, Xue Gong, Kelin Xia
Abstract: Recently, therapeutic peptides have demonstrated great promise for cancer treatment. To explore powerful anticancer peptides, artificial intelligence (AI)-based approaches have been developed to systematically screen potential candidates. However, the lack of efficient featurization of peptides has become a bottleneck for these machine-learning models. In this paper, we propose a topology-enhanced machine learning model (Top-ML) for anticancer peptides prediction. Our Top-ML employs peptide topological features derived from its sequence "connection" information characterized by vector and spectral descriptors. Our Top-ML model, employing an Extra-Trees classifier, has been validated on the AntiCP 2.0 and mACPpred 2.0 benchmark datasets, achieving state-of-the-art performance or results comparable to existing deep learning models, while providing greater interpretability. Our results highlight the potential of leveraging novel topology-based featurization to accelerate the identification of anticancer peptides.
Authors: Haiyan Zhao, Heng Zhao, Bo Shen, Ali Payani, Fan Yang, Mengnan Du
Abstract: Probing learned concepts in large language models (LLMs) is crucial for understanding how semantic knowledge is encoded internally. Training linear classifiers on probing tasks is a principle approach to denote the vector of a certain concept in the representation space. However, the single vector identified for a concept varies with both data and training, making it less robust and weakening its effectiveness in real-world applications. To address this challenge, we propose an approach to approximate the subspace representing a specific concept. Built on linear probing classifiers, we extend the concept vectors into Gaussian Concept Subspace (GCS). We demonstrate GCS's effectiveness through measuring its faithfulness and plausibility across multiple LLMs with different sizes and architectures. Additionally, we use representation intervention tasks to showcase its efficacy in real-world applications such as emotion steering. Experimental results indicate that GCS concept vectors have the potential to balance steering performance and maintaining the fluency in natural language generation tasks.
Authors: Collin Zhang, Tingwei Zhang, Vitaly Shmatikov
Abstract: We design, implement, and evaluate adversarial decoding, a new, generic text generation technique that produces readable documents for different adversarial objectives. Prior methods either produce easily detectable gibberish, or cannot handle objectives that include embedding similarity. In particular, they only work for direct attacks (such as jailbreaking) and cannot produce adversarial text for realistic indirect injection, e.g., documents that (1) are retrieved in RAG systems in response to broad classes of queries, and also (2) adversarially influence subsequent generation. We also show that fluency (low perplexity) is not sufficient to evade filtering. We measure the effectiveness of adversarial decoding for different objectives, including RAG poisoning, jailbreaking, and evasion of defensive filters, and demonstrate that it outperforms existing methods while producing readable adversarial documents.
Authors: Seonghwan Seo, Minsu Kim, Tony Shen, Martin Ester, Jinkyoo Park, Sungsoo Ahn, Woo Youn Kim
Abstract: Generative models in drug discovery have recently gained attention as efficient alternatives to brute-force virtual screening. However, most existing models do not account for synthesizability, limiting their practical use in real-world scenarios. In this paper, we propose RxnFlow, which sequentially assembles molecules using predefined molecular building blocks and chemical reaction templates to constrain the synthetic chemical pathway. We then train on this sequential generating process with the objective of generative flow networks (GFlowNets) to generate both highly rewarded and diverse molecules. To mitigate the large action space of synthetic pathways in GFlowNets, we implement a novel action space subsampling method. This enables RxnFlow to learn generative flows over extensive action spaces comprising combinations of 1.2 million building blocks and 71 reaction templates without significant computational overhead. Additionally, RxnFlow can employ modified or expanded action spaces for generation without retraining, allowing for the introduction of additional objectives or the incorporation of newly discovered building blocks. We experimentally demonstrate that RxnFlow outperforms existing reaction-based and fragment-based models in pocket-specific optimization across various target pockets. Furthermore, RxnFlow achieves state-of-the-art performance on CrossDocked2020 for pocket-conditional generation, with an average Vina score of -8.85 kcal/mol and 34.8% synthesizability.
Authors: Thomas Steinke, Milad Nasr, Arun Ganesh, Borja Balle, Christopher A. Choquette-Choo, Matthew Jagielski, Jamie Hayes, Abhradeep Guha Thakurta, Adam Smith, Andreas Terzis
Abstract: We propose a simple heuristic privacy analysis of noisy clipped stochastic gradient descent (DP-SGD) in the setting where only the last iterate is released and the intermediate iterates remain hidden. Namely, our heuristic assumes a linear structure for the model. We show experimentally that our heuristic is predictive of the outcome of privacy auditing applied to various training procedures. Thus it can be used prior to training as a rough estimate of the final privacy leakage. We also probe the limitations of our heuristic by providing some artificial counterexamples where it underestimates the privacy leakage. The standard composition-based privacy analysis of DP-SGD effectively assumes that the adversary has access to all intermediate iterates, which is often unrealistic. However, this analysis remains the state of the art in practice. While our heuristic does not replace a rigorous privacy analysis, it illustrates the large gap between the best theoretical upper bounds and the privacy auditing lower bounds and sets a target for further work to improve the theoretical privacy analyses. We also empirically support our heuristic and show existing privacy auditing attacks are bounded by our heuristic analysis in both vision and language tasks.
Authors: Esteban Garces Arias, Hannah Blocher, Julian Rodemann, Meimingwei Li, Christian Heumann, Matthias A{\ss}enmacher
Abstract: Open-ended text generation has become a prominent task in natural language processing due to the rise of powerful (large) language models. However, evaluating the quality of these models and the employed decoding strategies remains challenging because of trade-offs among widely used metrics such as coherence, diversity, and perplexity. Decoding methods often excel in some metrics while underperforming in others, complicating the establishment of a clear ranking. In this paper, we present novel ranking strategies within this multicriteria framework. Specifically, we employ benchmarking approaches based on partial orderings and present a new summary metric designed to balance existing automatic indicators, providing a more holistic evaluation of text generation quality. Our experiments demonstrate that the proposed methods offer a robust way to compare decoding strategies, and serve as valuable tools in guiding model selection for open-ended text generation tasks. Finally, we suggest future directions for improving evaluation methodologies in text generation. Our codebase, datasets, and models are publicly available.
Authors: Till Aczel, Roger Wattenhofer
Abstract: In lossy image compression, models face the challenge of either hallucinating details or generating out-of-distribution samples due to the information bottleneck. This implies that at times, introducing hallucinations is necessary to generate in-distribution samples. The optimal level of hallucination varies depending on image content, as humans are sensitive to small changes that alter the semantic meaning. We propose a novel compression method that dynamically balances the degree of hallucination based on content. We collect data and train a model to predict user preferences on hallucinations. By using this prediction to adjust the perceptual weight in the reconstruction loss, we develop a Conditionally Hallucinating compression model (ConHa) that outperforms state-of-the-art image compression methods. Code and images are available at https://polybox.ethz.ch/index.php/s/owS1k5JYs4KD4TA.
Authors: Zhenyu Jiang, Yuqi Xie, Kevin Lin, Zhenjia Xu, Weikang Wan, Ajay Mandlekar, Linxi Fan, Yuke Zhu
Abstract: Imitation learning from human demonstrations is an effective means to teach robots manipulation skills. But data acquisition is a major bottleneck in applying this paradigm more broadly, due to the amount of cost and human effort involved. There has been significant interest in imitation learning for bimanual dexterous robots, like humanoids. Unfortunately, data collection is even more challenging here due to the challenges of simultaneously controlling multiple arms and multi-fingered hands. Automated data generation in simulation is a compelling, scalable alternative to fuel this need for data. To this end, we introduce DexMimicGen, a large-scale automated data generation system that synthesizes trajectories from a handful of human demonstrations for humanoid robots with dexterous hands. We present a collection of simulation environments in the setting of bimanual dexterous manipulation, spanning a range of manipulation behaviors and different requirements for coordination among the two arms. We generate 21K demos across these tasks from just 60 source human demos and study the effect of several data generation and policy learning decisions on agent performance. Finally, we present a real-to-sim-to-real pipeline and deploy it on a real-world humanoid can sorting task. Generated datasets, simulation environments and additional results are at https://dexmimicgen.github.io/
Authors: Hoang H Nguyen, Khyati Mahajan, Vikas Yadav, Julian Salazar, Philip S. Yu, Masoud Hashemi, Rishabh Maheshwary
Abstract: Although multilingual LLMs have achieved remarkable performance across benchmarks, we find they continue to underperform on non-Latin script languages across contemporary LLM families. This discrepancy arises from the fact that LLMs are pretrained with orthographic scripts, which are dominated by Latin characters that obscure their shared phonology with non-Latin scripts. We propose leveraging phonemic transcriptions as complementary signals to induce script-invariant representations. Our study demonstrates that integrating phonemic signals improves performance across both non-Latin and Latin script languages, with a particularly significant impact on closing the performance gap between the two. Through detailed experiments, we show that phonemic and orthographic scripts retrieve distinct examples for in-context learning (ICL). This motivates our proposed Mixed-ICL retrieval strategy, where further aggregation from both leads to our significant performance improvements for both Latin script languages (up to 12.6%) and non-Latin script languages (up to 15.1%) compared to randomized ICL retrieval.
Authors: Shauli Ravfogel, Anej Svete, V\'esteinn Sn{\ae}bjarnarson, Ryan Cotterell
Abstract: Understanding and manipulating the causal generation mechanisms in language models is essential for controlling their behavior. Previous work has primarily relied on techniques such as representation surgery -- e.g., model ablations or manipulation of linear subspaces tied to specific concepts -- to \emph{intervene} on these models. To understand the impact of interventions precisely, it is useful to examine \emph{counterfactuals} -- e.g., how a given sentence would have appeared had it been generated by the model following a specific intervention. We highlight that counterfactual reasoning is conceptually distinct from interventions, as articulated in Pearl's causal hierarchy. Based on this observation, we propose a framework for generating true string counterfactuals by reformulating language models as a structural equation model using the Gumbel-max trick, which we called Gumbel counterfactual generation. This reformulation allows us to model the joint distribution over original strings and their counterfactuals resulting from the same instantiation of the sampling noise. We develop an algorithm based on hindsight Gumbel sampling that allows us to infer the latent noise variables and generate counterfactuals of observed strings. Our experiments demonstrate that the approach produces meaningful counterfactuals while at the same time showing that commonly used intervention techniques have considerable undesired side effects.
Authors: Laura Ruis, Maximilian Mozes, Juhan Bae, Siddhartha Rao Kamalakara, Dwarak Talupuru, Acyr Locatelli, Robert Kirk, Tim Rockt\"aschel, Edward Grefenstette, Max Bartolo
Abstract: The capabilities and limitations of Large Language Models have been sketched out in great detail in recent years, providing an intriguing yet conflicting picture. On the one hand, LLMs demonstrate a general ability to solve problems. On the other hand, they show surprising reasoning gaps when compared to humans, casting doubt on the robustness of their generalisation strategies. The sheer volume of data used in the design of LLMs has precluded us from applying the method traditionally used to measure generalisation: train-test set separation. To overcome this, we study what kind of generalisation strategies LLMs employ when performing reasoning tasks by investigating the pretraining data they rely on. For two models of different sizes (7B and 35B) and 2.5B of their pretraining tokens, we identify what documents influence the model outputs for three simple mathematical reasoning tasks and contrast this to the data that are influential for answering factual questions. We find that, while the models rely on mostly distinct sets of data for each factual question, a document often has a similar influence across different reasoning questions within the same task, indicating the presence of procedural knowledge. We further find that the answers to factual questions often show up in the most influential data. However, for reasoning questions the answers usually do not show up as highly influential, nor do the answers to the intermediate reasoning steps. When we characterise the top ranked documents for the reasoning questions qualitatively, we confirm that the influential documents often contain procedural knowledge, like demonstrating how to obtain a solution using formulae or code. Our findings indicate that the approach to reasoning the models use is unlike retrieval, and more like a generalisable strategy that synthesises procedural knowledge from documents doing a similar form of reasoning.
Authors: Zheng Ma, Zeping Mao, Ruixue Zhang, Jiazhen Chen, Lei Xin, Paul Shan, Ali Ghodsi, Ming Li
Abstract: Data-Independent Acquisition (DIA) was introduced to improve sensitivity to cover all peptides in a range rather than only sampling high-intensity peaks as in Data-Dependent Acquisition (DDA) mass spectrometry. However, it is not very clear how useful DIA data is for de novo peptide sequencing as the DIA data are marred with coeluted peptides, high noises, and varying data quality. We present a new deep learning method DIANovo, and address each of these difficulties, and improves the previous established system DeepNovo-DIA by from 25% to 81%, averaging 48%, for amino acid recall, and by from 27% to 89%, averaging 57%, for peptide recall, by equipping the model with a deeper understanding of coeluted DIA spectra. This paper also provides criteria about when DIA data could be used for de novo peptide sequencing and when not to by providing a comparison between DDA and DIA, in both de novo and database search mode. We find that while DIA excels with narrow isolation windows on older-generation instruments, it loses its advantage with wider windows. However, with Orbitrap Astral, DIA consistently outperforms DDA due to narrow window mode enabled. We also provide a theoretical explanation of this phenomenon, emphasizing the critical role of the signal-to-noise profile in the successful application of de novo sequencing.
Authors: Yifan Zhang
Abstract: The rapid advancement of large language models (LLMs) such as GPT-3, PaLM, and Llama has significantly transformed natural language processing, showcasing remarkable capabilities in understanding and generating language. However, these models often struggle with tasks requiring complex reasoning, particularly in mathematical problem-solving, due in part to the scarcity of large-scale, high-quality, domain-specific datasets necessary for training sophisticated reasoning abilities. To address this limitation, we introduce Template-based Data Generation (TDG), a novel approach that leverages LLMs (GPT-4) to automatically generate parameterized meta-templates, which are then used to synthesize a vast array of high-quality problems and solutions. Leveraging TDG, we create TemplateMath Part I: TemplateGSM, a dataset comprising over 7 million synthetically generated grade school math problems--each accompanied by code-based and natural language solutions--with the potential to generate an effectively unlimited number more. This dataset alleviates the scarcity of large-scale mathematical datasets and serves as a valuable resource for pre-training, fine-tuning, and evaluating LLMs in mathematical reasoning. Our method not only enables the generation of virtually infinite data but also elevates data augmentation to a new level by using GPT-4 for meta-template generation, ensuring diverse and high-quality problem structures. The TemplateMath Part I: TemplateGSM dataset is publicly available at https://huggingface.co/datasets/math-ai/TemplateGSM. The code is available at https://github.com/iiis-ai/TemplateMath.
URLs: https://huggingface.co/datasets/math-ai/TemplateGSM., https://github.com/iiis-ai/TemplateMath.
Authors: Taesung Kwon, Jong Chul Ye
Abstract: In this paper, we propose a novel framework for solving high-definition video inverse problems using latent image diffusion models. Building on recent advancements in spatio-temporal optimization for video inverse problems using image diffusion models, our approach leverages latent-space diffusion models to achieve enhanced video quality and resolution. To address the high computational demands of processing high-resolution frames, we introduce a pseudo-batch consistent sampling strategy, allowing efficient operation on a single GPU. Additionally, to improve temporal consistency, we present pseudo-batch inversion, an initialization technique that incorporates informative latents from the measurement. By integrating with SDXL, our framework achieves state-of-the-art video reconstruction across a wide range of spatio-temporal inverse problems, including complex combinations of frame averaging and various spatial degradations, such as deblurring, super-resolution, and inpainting. Unlike previous methods, our approach supports multiple aspect ratios (landscape, vertical, and square) and delivers HD-resolution reconstructions (exceeding 1280x720) in under 6 seconds per frame on a single NVIDIA 4090 GPU.
Authors: Songlin Yang, Jan Kautz, Ali Hatamizadeh
Abstract: Linear Transformers have gained attention as efficient alternatives to standard Transformers, but their performance in retrieval and long-context tasks has been limited. To address these limitations, recent work has explored two distinct mechanisms: gating for adaptive memory control and the delta update rule for precise memory modifications. We observe that these mechanisms are complementary: gating enables rapid memory erasure while the delta rule facilitates targeted updates. Building on this insight, we introduce the gated delta rule and develop a parallel training algorithm optimized for modern hardware. Our proposed architecture, Gated DeltaNet, consistently surpasses existing models like Mamba2 and DeltaNet across multiple benchmarks, including language modeling, common-sense reasoning, in-context retrieval, length extrapolation, and long-context understanding. We further enhance performance by developing hybrid architectures that combine Gated DeltaNet layers with sliding window attention or Mamba2 layers, achieving both improved training efficiency and superior task performance.
Authors: Tingting Ni, Maryam Kamgarpour
Abstract: We develop a model-free approach to optimally control stochastic, Markovian systems subject to a reach-avoid constraint. Specifically, the state trajectory must remain within a safe set while reaching a target set within a finite time horizon. Due to the time-dependent nature of these constraints, we show that, in general, the optimal policy for this constrained stochastic control problem is non-Markovian, which increases the computational complexity. To address this challenge, we apply the state-augmentation technique from arXiv:2402.19360, reformulating the problem as a constrained Markov decision process (CMDP) on an extended state space. This transformation allows us to search for a Markovian policy, avoiding the complexity of non-Markovian policies. To learn the optimal policy without a system model, and using only trajectory data, we develop a log-barrier policy gradient approach. We prove that under suitable assumptions, the policy parameters converge to the optimal parameters, while ensuring that the system trajectories satisfy the stochastic reach-avoid constraint with high probability.
Authors: Beibei Li (Deep Space Exploration Laboratory), Yutian Chi (Deep Space Exploration Laboratory), Yuming Wang (Deep Space Exploration Laboratory,School of Earth,Space Sciences University of Science,Technology of China)
Abstract: This study introduces a novel approach that integrates the magnetic field data correction from the Tianwen-1 Mars mission with a neural network architecture constrained by physical principles derived from Maxwell's equation equations. By employing a Transformer based model capable of efficiently handling sequential data, the method corrects measurement anomalies caused by satellite dynamics, instrument interference, and environmental noise. As a result, it significantly improves both the accuracy and the physical consistency of the calibrated data. Compared to traditional methods that require long data segments and manual intervention often taking weeks or even months to complete this new approach can finish calibration in just minutes to hours, and predictions are made within seconds. This innovation not only accelerates the process of space weather modeling and planetary magnetospheric studies but also provides a robust framework for future planetary exploration and solar wind interaction research.
Authors: Alexander Valverde, Luis Solano
Abstract: In Costa Rica, an average of 5 tons of seashells are extracted from ecosystems annually. Confiscated seashells, cannot be returned to their ecosystems due to the lack of origin recognition. To address this issue, we developed a convolutional neural network (CNN) specifically for seashell identification. We built a dataset from scratch, consisting of approximately 19000 images from the Pacific and Caribbean coasts. Using this dataset, the model achieved a classification accuracy exceeding 85%. The model has been integrated into a user-friendly application, which has classified over 36,000 seashells to date, delivering real-time results within 3 seconds per image. To further enhance the system's accuracy, an anomaly detection mechanism was incorporated to filter out irrelevant or anomalous inputs, ensuring only valid seashell images are processed.
Authors: Young Seok Jeon, Hongfei Yang, Huazhu Fu, Mengling Feng
Abstract: 3D models surpass 2D models in CT/MRI segmentation by effectively capturing inter-slice relationships. However, the added depth dimension substantially increases memory consumption. While patch-based training alleviates memory constraints, it significantly slows down the inference speed due to the sliding window (SW) approach. We propose No-More-Sliding-Window (NMSW), a novel end-to-end trainable framework that enhances the efficiency of generic 3D segmentation backbone during an inference step by eliminating the need for SW. NMSW employs a differentiable Top-k module to selectively sample only the most relevant patches, thereby minimizing redundant computations. When patch-level predictions are insufficient, the framework intelligently leverages coarse global predictions to refine results. Evaluated across 3 tasks using 3 segmentation backbones, NMSW achieves competitive accuracy compared to SW inference while significantly reducing computational complexity by 91% (88.0 to 8.00 TMACs). Moreover, it delivers a 9.1x faster inference on the H100 GPU (99.0 to 8.3 sec) and a 11.1x faster inference on the Xeon Gold CPU (2110 to 189 sec). NMSW is model-agnostic, further boosting efficiency when integrated with any existing efficient segmentation backbones.
Authors: Aayush Mishra, Daniel Habermann, Marvin Schmitt, Stefan T. Radev, Paul-Christian B\"urkner
Abstract: Neural amortized Bayesian inference (ABI) can solve probabilistic inverse problems orders of magnitude faster than classical methods. However, neural ABI is not yet sufficiently robust for widespread and safe applicability. In particular, when performing inference on observations outside of the scope of the simulated data seen during training, for example, because of model misspecification, the posterior approximations are likely to become highly biased. Due to the bad pre-asymptotic behavior of current neural posterior estimators in the out-of-simulation regime, the resulting estimation biases cannot be fixed in acceptable time by just simulating more training data. In this proof-of-concept paper, we propose a semi-supervised approach that enables training not only on (labeled) simulated data generated from the model, but also on unlabeled data originating from any source, including real-world data. To achieve the latter, we exploit Bayesian self-consistency properties that can be transformed into strictly proper losses without requiring knowledge of true parameter values, that is, without requiring data labels. The results of our initial experiments show remarkable improvements in the robustness of ABI on out-of-simulation data. Even if the observed data is far away from both labeled and unlabeled training data, inference remains highly accurate. If our findings also generalize to other scenarios and model classes, we believe that our new method represents a major breakthrough in neural ABI.
Authors: Shivam Singh, Karthik Swaminathan, Nabanita Dash, Ramandeep Singh, Snehasis Banerjee, Mohan Sridharan, Madhava Krishna
Abstract: An embodied agent assisting humans is often asked to complete new tasks, and there may not be sufficient time or labeled examples to train the agent to perform these new tasks. Large Language Models (LLMs) trained on considerable knowledge across many domains can be used to predict a sequence of abstract actions for completing such tasks, although the agent may not be able to execute this sequence due to task-, agent-, or domain-specific constraints. Our framework addresses these challenges by leveraging the generic predictions provided by LLM and the prior domain knowledge encoded in a Knowledge Graph (KG), enabling an agent to quickly adapt to new tasks. The robot also solicits and uses human input as needed to refine its existing knowledge. Based on experimental evaluation in the context of cooking and cleaning tasks in simulation domains, we demonstrate that the interplay between LLM, KG, and human input leads to substantial performance gains compared with just using the LLM. Project website{\S}: https://sssshivvvv.github.io/adaptbot/
Authors: Zhifei Yang, Keyang Lu, Chao Zhang, Jiaxing Qi, Hanqi Jiang, Ruifei Ma, Shenglin Yin, Yifan Xu, Mingzhe Xing, Zhen Xiao, Jieyi Long, Xiangde Liu, Guangyao Zhai
Abstract: Controllable 3D scene generation has extensive applications in virtual reality and interior design, where the generated scenes should exhibit high levels of realism and controllability in terms of geometry. Scene graphs provide a suitable data representation that facilitates these applications. However, current graph-based methods for scene generation are constrained to text-based inputs and exhibit insufficient adaptability to flexible user inputs, hindering the ability to precisely control object geometry. To address this issue, we propose MMGDreamer, a dual-branch diffusion model for scene generation that incorporates a novel Mixed-Modality Graph, visual enhancement module, and relation predictor. The mixed-modality graph allows object nodes to integrate textual and visual modalities, with optional relationships between nodes. It enhances adaptability to flexible user inputs and enables meticulous control over the geometry of objects in the generated scenes. The visual enhancement module enriches the visual fidelity of text-only nodes by constructing visual representations using text embeddings. Furthermore, our relation predictor leverages node representations to infer absent relationships between nodes, resulting in more coherent scene layouts. Extensive experimental results demonstrate that MMGDreamer exhibits superior control of object geometry, achieving state-of-the-art scene generation performance. Project page: https://yangzhifeio.github.io/project/MMGDreamer.
Authors: Yingce Xia, Peiran Jin, Shufang Xie, Liang He, Chuan Cao, Renqian Luo, Guoqing Liu, Yue Wang, Zequn Liu, Yuan-Jyue Chen, Zekun Guo, Yeqi Bai, Pan Deng, Yaosen Min, Ziheng Lu, Hongxia Hao, Han Yang, Jielan Li, Chang Liu, Jia Zhang, Jianwei Zhu, Ran Bi, Kehan Wu, Wei Zhang, Kaiyuan Gao, Qizhi Pei, Qian Wang, Xixian Liu, Yanting Li, Houtian Zhu, Yeqing Lu, Mingqian Ma, Zun Wang, Tian Xie, Krzysztof Maziarz, Marwin Segler, Zhao Yang, Zilong Chen, Yu Shi, Shuxin Zheng, Lijun Wu, Chen Hu, Peggy Dai, Tie-Yan Liu, Haiguang Liu, Tao Qin
Abstract: Foundation models have revolutionized natural language processing and artificial intelligence, significantly enhancing how machines comprehend and generate human languages. Inspired by the success of these foundation models, researchers have developed foundation models for individual scientific domains, including small molecules, materials, proteins, DNA, RNA and even cells. However, these models are typically trained in isolation, lacking the ability to integrate across different scientific domains. Recognizing that entities within these domains can all be represented as sequences, which together form the "language of nature", we introduce Nature Language Model (NatureLM), a sequence-based science foundation model designed for scientific discovery. Pre-trained with data from multiple scientific domains, NatureLM offers a unified, versatile model that enables various applications including: (i) generating and optimizing small molecules, proteins, RNA, and materials using text instructions; (ii) cross-domain generation/design, such as protein-to-molecule and protein-to-RNA generation; and (iii) top performance across different domains, matching or surpassing state-of-the-art specialist models. NatureLM offers a promising generalist approach for various scientific tasks, including drug discovery (hit generation/optimization, ADMET optimization, synthesis), novel material design, and the development of therapeutic proteins or nucleotides. We have developed NatureLM models in different sizes (1 billion, 8 billion, and 46.7 billion parameters) and observed a clear improvement in performance as the model size increases.
Authors: Xiaoya Lu, Dongrui Liu, Yi Yu, Luxin Xu, Jing Shao
Abstract: Despite the rapid development of safety alignment techniques for LLMs, defending against multi-turn jailbreaks is still a challenging task. In this paper, we conduct a comprehensive comparison, revealing that some existing defense methods can improve the robustness of LLMs against multi-turn jailbreaks but compromise usability, i.e., reducing general capabilities or causing the over-refusal problem. From the perspective of mechanism interpretability of LLMs, we discover that these methods fail to establish a boundary that exactly distinguishes safe and harmful feature representations. Therefore, boundary-safe representations close to harmful representations are inevitably disrupted, leading to a decline in usability. To address this issue, we propose X-Boundary to push harmful representations away from boundary-safe representations and obtain an exact distinction boundary. In this way, harmful representations can be precisely erased without disrupting safe ones. Experimental results show that X-Boundary achieves state-of-the-art defense performance against multi-turn jailbreaks, while reducing the over-refusal rate by about 20% and maintaining nearly complete general capability. Furthermore, we theoretically prove and empirically verify that X-Boundary can accelerate the convergence process during training. Please see our code at: https://github.com/AI45Lab/X-Boundary.
Authors: Sujan Sai Gannamaneni, Rohil Prakash Rao, Michael Mock, Maram Akila, Stefan Wrobel
Abstract: Slice discovery methods (SDMs) are prominent algorithms for finding systematic weaknesses in DNNs. They identify top-k semantically coherent slices/subsets of data where a DNN-under-test has low performance. For being directly useful, slices should be aligned with human-understandable and relevant dimensions, which, for example, are defined by safety and domain experts as part of the operational design domain (ODD). While SDMs can be applied effectively on structured data, their application on image data is complicated by the lack of semantic metadata. To address these issues, we present an algorithm that combines foundation models for zero-shot image classification to generate semantic metadata with methods for combinatorial search to find systematic weaknesses in images. In contrast to existing approaches, ours identifies weak slices that are in line with pre-defined human-understandable dimensions. As the algorithm includes foundation models, its intermediate and final results may not always be exact. Therefore, we include an approach to address the impact of noisy metadata. We validate our algorithm on both synthetic and real-world datasets, demonstrating its ability to recover human-understandable systematic weaknesses. Furthermore, using our approach, we identify systematic weaknesses of multiple pre-trained and publicly available state-of-the-art computer vision DNNs.
Authors: Wei Dai, Jun Liu
Abstract: Efficient evaluation of three-dimensional (3D) medical images is crucial for diagnostic and therapeutic practices in healthcare. Recent years have seen a substantial uptake in applying deep learning and computer vision to analyse and interpret medical images. Traditional approaches, such as convolutional neural networks (CNNs) and vision transformers (ViTs), face significant computational challenges, prompting the need for architectural advancements. Recent efforts have led to the introduction of novel architectures like the ``Mamba'' model as alternative solutions to traditional CNNs or ViTs. The Mamba model excels in the linear processing of one-dimensional data with low computational demands. However, Mamba's potential for 3D medical image analysis remains underexplored and could face significant computational challenges as the dimension increases. This manuscript presents MobileViM, a streamlined architecture for efficient segmentation of 3D medical images. In the MobileViM network, we invent a new dimension-independent mechanism and a dual-direction traversing approach to incorporate with a vision-Mamba-based framework. MobileViM also features a cross-scale bridging technique to improve efficiency and accuracy across various medical imaging modalities. With these enhancements, MobileViM achieves segmentation speeds exceeding 90 frames per second (FPS) on a single graphics processing unit (i.e., NVIDIA RTX 4090). This performance is over 24 FPS faster than the state-of-the-art deep learning models for processing 3D images with the same computational resources. In addition, experimental evaluations demonstrate that MobileViM delivers superior performance, with Dice similarity scores reaching 92.72%, 86.69%, 80.46%, and 77.43% for PENGWIN, BraTS2024, ATLAS, and Toothfairy2 datasets, respectively, which significantly surpasses existing models.
Authors: Ruida Hu, Chao Peng, Xinchen Wang, Cuiyun Gao
Abstract: Environment configuration is a critical yet time-consuming step in software development, especially when dealing with unfamiliar code repositories. While Large Language Models (LLMs) demonstrate the potential to accomplish software engineering tasks, existing methods for environment configuration often rely on manual efforts or fragile scripts, leading to inefficiencies and unreliable outcomes. We introduce Repo2Run, the first LLM-based agent designed to fully automate environment configuration and generate executable Dockerfiles for arbitrary Python repositories. We address two major challenges: (1) enabling the LLM agent to configure environments within isolated Docker containers, and (2) ensuring the successful configuration process is recorded and accurately transferred to a Dockerfile without error. To achieve this, we propose atomic configuration synthesis, featuring a dual-environment architecture (internal and external environment) with a rollback mechanism to prevent environment "pollution" from failed commands, guaranteeing atomic execution (execute fully or not at all) and a Dockerfile generator to transfer successful configuration steps into runnable Dockerfiles. We evaluate Repo2Run~on our proposed benchmark of 420 recent Python repositories with unit tests, where it achieves an 86.0% success rate, outperforming the best baseline by 63.9%. Repo2Run is available at https://github.com/bytedance/Repo2Run.
Authors: Yi Xu, Laura Ruis, Tim Rockt\"aschel, Robert Kirk
Abstract: Automatic evaluation methods based on large language models (LLMs) are emerging as the standard tool for assessing the instruction-following abilities of LLM-based agents. The most common method in this paradigm, pairwise comparisons with a baseline model, critically depends on the assumption of transitive preferences. However, the validity of this assumption remains largely unexplored. In this study, we investigate the presence of non-transitivity within the AlpacaEval framework and analyze its effects on model rankings. We find that LLM judges exhibit non-transitive preferences, leading to rankings that are sensitive to the choice of the baseline model. To mitigate this issue, we show that round-robin tournaments combined with Bradley-Terry models of preference can produce more reliable rankings. Notably, our method increases both the Spearman correlation and the Kendall correlation with Chatbot Arena (95.0% -> 96.4% and 82.1% -> 86.3% respectively). To address the computational cost of round-robin tournaments, we propose Swiss-Wise Iterative Matchmaking (Swim) tournaments, using a dynamic matching strategy to capture the benefits of round-robin tournaments while maintaining computational efficiency.
Authors: Leena Mathur, Marian Qian, Paul Pu Liang, Louis-Philippe Morency
Abstract: Social reasoning abilities are crucial for AI systems to effectively interpret and respond to multimodal human communication and interaction within social contexts. We introduce Social Genome, the first benchmark for fine-grained, grounded social reasoning abilities of multimodal models. Social Genome contains 272 videos of interactions and 1,486 human-annotated reasoning traces related to inferences about these interactions. These traces contain 5,777 reasoning steps that reference evidence from visual cues, verbal cues, vocal cues, and external knowledge (contextual knowledge external to videos). Social Genome is also the first modeling challenge to study external knowledge in social reasoning. Social Genome computes metrics to holistically evaluate semantic and structural qualities of model-generated social reasoning traces. We demonstrate the utility of Social Genome through experiments with state-of-the-art models, identifying performance gaps and opportunities for future research to improve the grounded social reasoning abilities of multimodal models.
Authors: Sophie Xhonneux, David Dobre, Mehrnaz Mofakhami, Leo Schwinn, Gauthier Gidel
Abstract: Most safety training methods for large language models (LLMs) based on fine-tuning rely on dramatically changing the output distribution of the model when faced with a harmful request, shifting it from an unsafe answer to a refusal to respond. These methods inherently compromise model capabilities and might make auto-regressive models vulnerable to attacks that make likely an initial token of affirmative response. To avoid that, we propose to expand the model's vocabulary with a special token we call red flag token (
Authors: Thanh Trung Vu, Andreas Kofler, Kostas Papafitsoros
Abstract: We extend a recently introduced deep unrolling framework for learning spatially varying regularisation parameters in inverse imaging problems to the case of Total Generalised Variation (TGV). The framework combines a deep convolutional neural network (CNN) inferring the two spatially varying TGV parameters with an unrolled algorithmic scheme that solves the corresponding variational problem. The two subnetworks are jointly trained end-to-end in a supervised fashion and as such the CNN learns to compute those parameters that drive the reconstructed images as close to the ground truth as possible. Numerical results in image denoising and MRI reconstruction show a significant qualitative and quantitative improvement compared to the best TGV scalar parameter case as well as to other approaches employing spatially varying parameters computed by unsupervised methods. We also observe that the inferred spatially varying parameter maps have a consistent structure near the image edges, asking for further theoretical investigations. In particular, the parameter that weighs the first-order TGV term has a triple-edge structure with alternating high-low-high values whereas the one that weighs the second-order term attains small values in a large neighbourhood around the edges.
Authors: Yijia Xiao, Wanjia Zhao, Junkai Zhang, Yiqiao Jin, Han Zhang, Zhicheng Ren, Renliang Sun, Haixin Wang, Guancheng Wan, Pan Lu, Xiao Luo, Yu Zhang, James Zou, Yizhou Sun, Wei Wang
Abstract: Protein-specific large language models (Protein LLMs) are revolutionizing protein science by enabling more efficient protein structure prediction, function annotation, and design. While existing surveys focus on specific aspects or applications, this work provides the first comprehensive overview of Protein LLMs, covering their architectures, training datasets, evaluation metrics, and diverse applications. Through a systematic analysis of over 100 articles, we propose a structured taxonomy of state-of-the-art Protein LLMs, analyze how they leverage large-scale protein sequence data for improved accuracy, and explore their potential in advancing protein engineering and biomedical research. Additionally, we discuss key challenges and future directions, positioning Protein LLMs as essential tools for scientific discovery in protein science. Resources are maintained at https://github.com/Yijia-Xiao/Protein-LLM-Survey.
Authors: Hengzhi He, Shirong Xu, Guang Cheng
Abstract: Recent studies identified an intriguing phenomenon in recursive generative model training known as model collapse, where models trained on data generated by previous models exhibit severe performance degradation. Addressing this issue and developing more effective training strategies have become central challenges in generative model research. In this paper, we investigate this phenomenon theoretically within a novel framework, where generative models are iteratively trained on a combination of newly collected real data and synthetic data from the previous training step. To develop an optimal training strategy for integrating real and synthetic data, we evaluate the performance of a weighted training scheme in various scenarios, including Gaussian distribution estimation and linear regression. We theoretically characterize the impact of the mixing proportion and weighting scheme of synthetic data on the final model's performance. Our key finding is that, across different settings, the optimal weighting scheme under different proportions of synthetic data asymptotically follows a unified expression, revealing a fundamental trade-off between leveraging synthetic data and generative model performance. Notably, in some cases, the optimal weight assigned to real data corresponds to the reciprocal of the golden ratio. Finally, we validate our theoretical results on extensive simulated datasets and a real tabular dataset.
Authors: Sijin Sun, Ming Deng, Xingrui Yu, Xinyu Xi, Liangbin Zhao
Abstract: Metal defect detection is critical in industrial quality assurance, yet existing methods struggle with grayscale variations and complex defect states, limiting its robustness. To address these challenges, this paper proposes a Self-Adaptive Gamma Context-Aware SSM-based model(GCM-DET). This advanced detection framework integrating a Dynamic Gamma Correction (GC) module to enhance grayscale representation and optimize feature extraction for precise defect reconstruction. A State-Space Search Management (SSM) architecture captures robust multi-scale features, effectively handling defects of varying shapes and scales. Focal Loss is employed to mitigate class imbalance and refine detection accuracy. Additionally, the CD5-DET dataset is introduced, specifically designed for port container maintenance, featuring significant grayscale variations and intricate defect patterns. Experimental results demonstrate that the proposed model achieves substantial improvements, with mAP@0.5 gains of 27.6\%, 6.6\%, and 2.6\% on the CD5-DET, NEU-DET, and GC10-DET datasets.
Authors: Hang Yu, Takayuki Miyagi
Abstract: Constructing fast and accurate surrogate models is a key ingredient for making robust predictions in many topics. We introduce a new model, the Multiparameter Eigenvalue Problem (MEP) emulator. The new method connects emulators and can make predictions directly from observables to observables. We present that the MEP emulator can be trained with data from Eigenvector Continuation (EC) and Parametric Matrix Model (PMM) emulators. A simple simulation on a one-dimensional lattice confirms the performance of the MEP emulator. Using $^{28}$O as an example, we also demonstrate that the predictive probability distribution of the target observables can be easily obtained through the new emulator.
Authors: Mohammad Albinhassan, Pranava Madhyastha, Alessandra Russo
Abstract: Ensuring both syntactic and semantic correctness in Large Language Model (LLM) outputs remains a significant challenge, despite being critical for real-world deployment. In this paper, we introduce $\texttt{SEM-CTRL}$, a unified approach that enforces rich context-sensitive constraints and task- and instance-specific semantics directly on an LLM decoder. Our approach integrates token-level MCTS, which is guided by specific syntactic and semantic constraints. The constraints over the desired outputs are expressed using Answer Set Grammars -- a logic-based formalism that generalizes context-sensitive grammars while incorporating background knowledge to represent task-specific semantics. We show that our approach guarantees correct completions for any off-the-shelf LLM without the need for fine-tuning. We evaluate $\texttt{SEM-CTRL}$ on a range of tasks, including synthetic grammar synthesis, combinatorial reasoning, and planning. Our results demonstrate that $\texttt{SEM-CTRL}$ allows small pre-trained LLMs to efficiently outperform larger variants and state-of-the-art reasoning models (e.g., o1-preview) while simultaneously guaranteeing solution correctness.
Authors: Khoi Anh Nguyen, Linh Yen Vu, Thang Dinh Duong, Thuan Nguyen Duong, Huy Thanh Nguyen, Vinh Quang Dinh
Abstract: Visual Question Answering (VQA) is a multimodal task requiring reasoning across textual and visual inputs, which becomes particularly challenging in low-resource languages like Vietnamese due to linguistic variability and the lack of high-quality datasets. Traditional methods often rely heavily on extensive annotated datasets, computationally expensive pipelines, and large pre-trained models, specifically in the domain of Vietnamese VQA, limiting their applicability in such scenarios. To address these limitations, we propose a training framework that combines a paraphrase-based feature augmentation module with a dynamic curriculum learning strategy. Explicitly, augmented samples are considered "easy" while raw samples are regarded as "hard". The framework then utilizes a mechanism that dynamically adjusts the ratio of easy to hard samples during training, progressively modifying the same dataset to increase its difficulty level. By enabling gradual adaptation to task complexity, this approach helps the Vietnamese VQA model generalize well, thus improving overall performance. Experimental results show consistent improvements on the OpenViVQA dataset and mixed outcomes on the ViVQA dataset, highlighting both the potential and challenges of our approach in advancing VQA for Vietnamese language.
Authors: Ting-Wei Liao, Chih-Hsun Lin, Yu-Lin Tsai, Takao Murakami, Chia-Mu Yu, Jun Sakuma, Chun-Ying Huang, Hiroaki Kikuchi
Abstract: Local Differential Privacy (LDP) has been widely adopted to protect user privacy in decentralized data collection. However, recent studies have revealed that LDP protocols are vulnerable to data poisoning attacks, where malicious users manipulate their reported data to distort aggregated results. In this work, we present the first study on data poisoning attacks targeting LDP range query protocols, focusing on both tree-based and grid-based approaches. We identify three key challenges in executing such attacks, including crafting consistent and effective fake data, maintaining data consistency across levels or grids, and preventing server detection. To address the first two challenges, we propose novel attack methods that are provably optimal, including a tree-based attack and a grid-based attack, designed to manipulate range query results with high effectiveness. \textbf{Our key finding is that the common post-processing procedure, Norm-Sub, in LDP range query protocols can help the attacker massively amplify their attack effectiveness.} In addition, we study a potential countermeasure, but also propose an adaptive attack capable of evading this defense to address the third challenge. We evaluate our methods through theoretical analysis and extensive experiments on synthetic and real-world datasets. Our results show that the proposed attacks can significantly amplify estimations for arbitrary range queries by manipulating a small fraction of users, providing 5-10x more influence than a normal user to the estimation.