Machine Learning
Global Optimality for Constrained Exploration via Penalty Regularization
Florian Wolf, Ilyas Fatkhullin and Niao He
Efficient exploration is a central problem in reinforcement learning and is often formalized as maximizing the
entropy of the state-action occupancy measure. While unconstrained maximum-
entropy exploration is relatively well understood, real-world exploration is often constrained by safety, resource, or imitation requirements. This constrained setting is particularly challenging because
entropy maximization lacks additive structure, rendering Bellman-equation-based methods inapplicable. Moreover, scalable approaches require policy parameterization, inducing non-convexity in both the objective and the constraints. To our knowledge, the only prior model-free policy-gradient approach for this setting under general policy parameterization is due to Ying et al. (2025). Unfortunately, their guarantees are limited to weak regret and ergodic averages, which do not imply that the final output is a single deployable policy that is near-optimal and nearly feasible. In this work we take a different approach to this problem, and propose Policy Gradient Penalty (PGP) method, a single-loop policy-space method that enforces general convex occupancy-measure constraints via quadratic-penalty regularization. PGP constructs pseudo-rewards that yield gradient estimates of the penalized objective, subsequently exploiting the classical Policy Gradient
Theorem. We further establish the regularity of the penalized objective, providing the smoothness properties needed to justify the convergence of PGP. Leveraging hidden convexity and strong duality, we then establish global last-iterate convergence guarantees, attaining an $\epsilon$-optimal constrained
entropy value with $\epsilon$-bounded constraint violation despite policy-induced non-convexity. We validate PGP through ablations on a grid-world
benchmark and further demonstrate scalability on two challenging continuous-control tasks.
Read the full abstract ▾
Machine Learning · Cross-listed from cs.IR
Efficient Multivector Retrieval with Token-Aware Clustering and Hierarchical Indexing
Silvio Martinico, Franco Maria Nardini, Cosimo Rulli et al. (+1 more)
Multivector
retrieval models achieve state-of-the-art effectiveness through fine-grained
token-level representations, but their deployment incurs substantial computational and
memory costs. Current solutions, based on the well-known k-means clustering
algorithm, group similar vectors together to enable both effective compression and efficient
retrieval. However, standard k-means scales poorly with the number of clusters and
dataset size, and favours frequent
tokens during training while underrepresenting rare, discriminative ones. In this work, we introduce TACHIOM, a multivector
retrieval system that exploits
token-level structure to significantly accelerate both clustering and
retrieval. By accounting for
tokens' distribution during centroid allocation, TACHIOM easily scales to millions of centroids, enabling highly accurate document scoring using only centroids, avoiding expensive
token-level computation. TACHIOM combines a graph-based index over centroids with an optimized Product Quantization layout for efficient final scoring. Experiments on MS-MARCOv1 and LoTTE show that TACHIOM achieves up to $247\times$ faster clustering than k-means and up to $9.8\times$
retrieval speedup over state-of-the-art systems while maintaining comparable or superior effectiveness.
Read the full abstract ▾
All authors Silvio Martinico, Franco Maria Nardini, Cosimo Rulli, Rossano Venturini
Machine Learning · Cross-listed from cs.CV
Beyond Gaussian Bottlenecks: Topologically Aligned Encoding of Vision-Transformer Feature Spaces
Andrew Bond, Ilkin Umut Melanlioglu, Erkut Erdem et al. (+1 more)
Modern visual world modeling systems increasingly rely on high-capacity architectures and large-scale data to produce plausible motion, yet they often fail to preserve underlying 3D geometry or physically consistent camera dynamics. A key limitation lies not only in model capacity, but in the latent representations used to encode
geometric structure. We propose S$^2$VAE, a geometry-first latent learning framework that focuses on compressing and representing the latent 3D state of a scene, including camera motion, depth, and point-level structure, rather than modeling appearance alone. Building on representations from a Visual Geometry Grounded
Transformer (VGGT), we introduce a novel type of variational autoencoder using a product of Power Spherical latent distributions, explicitly enforcing hyperspherical structure in the bottleneck to preserve directional and
geometric semantics under strong compression. Across depth estimation, camera pose recovery, and point cloud reconstruction, we show that geometry-aligned hyperspherical latents consistently outperform conventional Gaussian bottlenecks, particularly in high-compression regimes. Our results highlight latent geometry as a first-class design choice for physically grounded visual and world models.
Read the full abstract ▾
All authors Andrew Bond, Ilkin Umut Melanlioglu, Erkut Erdem, Aykut Erdem
Machine Learning
Do Sparse Autoencoders Capture Concept Manifolds?
Usha Bhalla, Thomas Fel, Can Rager et al. (+9 more)
Sparse autoencoders (SAEs) are widely used to extract interpretable features from neural network representations, often under the implicit assumption that concepts correspond to independent linear directions. However, a growing body of evidence suggests that many concepts are instead organized along low-dimensional
manifolds encoding continuous
geometric relationships. This raises three basic questions: what does it mean for an SAE to capture a
manifold, when do existing SAE architectures do so, and how? We develop a theoretical framework that answers these questions and show that SAEs can capture
manifolds in two fundamentally different ways: globally, by allocating a compact group of atoms whose linear span contains the entire
manifold, or locally, by distributing it across features that each selectively tile a restricted region of the underlying geometry. Empirically, we find that SAEs suboptimally recover continuous structures, mixing the global subspace and local tiling solutions in a fragmented regime we call dilution. This explains why
manifold structure is rarely visible at the level of individual concepts and motivates post-hoc unsupervised discovery methods that search for coherent groups of atoms rather than isolated directions. More broadly, our results suggest that future representation learning methods should treat
geometric objects, not just individual directions, as the basic units of interpretability.
Read the full abstract ▾
All authors Usha Bhalla, Thomas Fel, Can Rager, Sheridan Feucht, Tal Haklay, Daniel Wurgaft, Siddharth Boppana, Matthew Kowal, Vasudev Shyam, Jack Merullo, Atticus Geiger, Ekdeep Singh Lubana
Machine Learning · Cross-listed from cs.SE
Hierarchical Fault Detection and Diagnosis for Transformer Architectures
Sigma Jahan, Saurabh Singh Rajput, Tushar Sharma et al. (+1 more)
Transformers now underpin critical AI systems across industry and research. Yet their faults can silently alter model behavior without runtime errors, and existing techniques offer little support for tracing these failures to their component and root cause. Such faults evade detection because loss and numerical values stay normal, and the visible symptom rarely identifies the component responsible. We present DEFault++, a hierarchical learning-based technique that first detects a fault, then identifies the affected component, and finally the cause within it, helping developers effectively debug
transformer models. DEFault++ organizes component-level runtime measurements with a Fault Propagation Graph (FPG), a structural prior over the architecture's dependency paths, and reports the evidence behind each diagnosis. To train and evaluate it, we construct DEFault-bench, a
benchmark of 5,556 labeled runs from mutation testing across seven models, nine tasks, and both encoder and decoder architectures. DEFault++ improves fault detection over four prior techniques, reaching an F1 of 0.826–0.909, and in a developer study with 21 participants, it raises repair accuracy from 57.1% to 83.3%. These results show that
transformer fault diagnosis benefits from component-level measurements and architecture-aware
reasoning rather than model-level behavior alone, and DEFault-bench provides a foundation for further research on
transformer fault diagnosis.
Read the full abstract ▾
All authors Sigma Jahan, Saurabh Singh Rajput, Tushar Sharma, Mohammad Masudur Rahman
Machine Learning
Auto-FlexSwitch: Efficient Dynamic Model Merging via Learnable Task Vector Compression
Junqi Gao, Dazhi Zhang, Zhichang Guo et al. (+3 more)
Model merging has attracted
attention as an effective path toward multi-task adaptation by integrating knowledge from multiple task-specific models. Among existing approaches, dynamic merging mitigates performance degradation caused by conflicting parameter updates across tasks by flexibly combining task-specific parameters at inference time, thereby maintaining high performance. However, these methods require storing independent parameters for each task, resulting in prohibitive storage overhead. To address this issue, we first experimentally demonstrate that the fine-tuned weight increments (referred to as task vectors) exhibit an impulse-like activation pattern and high robustness to low-bit representations. Driven by this insight, we propose T-Switch, which decomposes task vectors into three compact components: a binary sparse mask, a sign vector, and a scalar scaling factor, achieving high-fidelity approximation at high compression ratios. We then introduce Auto-Switch, a training-free merging scheme that automatically composes task vectors via feature similarity
retrieval. Building on this, we develop Auto-Switch, a training-free merging scheme that automatically assembles task vectors through feature similarity
retrieval. Furthermore, to transform task vector sparsification and quantization from static rules to adaptive learning, we propose FlexSwitch, a learnable framework which jointly optimizes the compression strategy for each model unit via Learnable Gating Sparsification (LGS) and Bit-width Adaptive Selection (BAS), while employing the Sparsity-Aware Storage Strategy (SASS) to select the optimal storage encoding structure. Finally, by incorporating a K-Nearest Neighbor (KNN) inference scheme with a learnable low-rank metric, we present Auto-FlexSwitch, a dynamic model merging approach that supports highly efficient task vector compression.
Read the full abstract ▾
All authors Junqi Gao, Dazhi Zhang, Zhichang Guo, Biqing Qi, Yi Ran, Wangmeng Zuo
Machine Learning
Neural Aided Kalman Filtering for UAV State Estimation in Degraded Sensing Environments
Akhil Gupta and Erhan Guven
Accurate state estimation of nonlinear dynamical systems is fundamental to modern aerospace operations across air, sea, and space domains. Online tracking of
adversarial unmanned aerial vehicles (
UAVs) is especially challenging due to agile nonlinear motion, noisy and sparse sensor measurements, and unknown control inputs; conditions that violate key assumptions of classical Kalman filter variants and degrade estimation performance. Neural networks (NNs) can learn complex nonlinear relationships from data, but lack principled uncertainty quantification, which is critical for state estimation tasks where confidence bounds drive downstream decisions. We address this with
Bayesian Neural Networks (BNNs), which model uncertainty through distributions over network weights and produce predictive means and uncertainties via Monte Carlo sampling. Building on this, we propose the
Bayesian Neural Kalman Filter (BNKF): a hybrid framework coupling a trained BNN with a Kalman correction step for robust online
UAV state estimation. Unlike related neural Kalman approaches, BNKF produces full state predictions and incorporates
Bayesian uncertainty directly into covariance propagation, improving robustness under high noise conditions. We evaluate BNKF under varying radar noise levels and sampling rates using synthetic nonlinear
UAV flight data. Five fold cross validation demonstrates that BNKF outperforms Extended and Unscented Kalman Filters in accuracy, precision, and truth containment under degraded sensing. An ensemble variant (BNKFe) further improves precision in high-noise edge cases at a slight accuracy tradeoff. Runtime analysis confirms minimal inference overhead, supporting real-time deployment feasibility.
Read the full abstract ▾
Machine Learning
FiLMMeD: Feature-wise Linear Modulation for Cross-Problem Multi-Depot Vehicle Routing
Arthur Corrêa, Paulo Nascimento and Samuel Moniz
Solving practical multi-depot vehicle routing problems (MDVRP) is a challenging optimization task central to modern logistics, increasingly driven by e-commerce. To address the MDVRP's computational complexity, neural-based combinatorial optimization methods offer a promising scalable alternative to traditional approaches. However, neural-based methods typically rely on rigid architectures and input encodings tailored to specific problem formulations. In real-world settings, heterogeneous constraints create multiple MDVRP variants, limiting the applicability of such models. While multi-task learning (MTL) has begun to accelerate the development of unified neural-based solvers, prior works focus almost exclusively on single-depot VRPs, leaving the MDVRP unaddressed. To bridge this gap, we propose Feature-wise Linear Modulation for Cross-Problem Multi-Depot Vehicle Routing (FiLMMeD), a novel unified neural-based model for 24 different MDVRP variants. We introduce three main contributions: (1) to improve the model's generalization, we augment the standard
Transformer encoder with Feature-wise Linear Modulation (FiLM), which dynamically conditions learned internal representations based on the active set of constraints; (2) we provide an initial demonstration of Preference Optimization in the MTL setting, establishing it as a superior alternative to Reinforcement Learning for future MTL works; (3) to mitigate the generalization gap caused by the introduction of multi-depot constraints, we introduce a targeted curriculum learning strategy that progressively exposes the model to increasingly more complex constraint interactions. Extensive experiments on 24 MDVRP variants (including 8 novel formulations) and 16 single-depot VRPs confirm the effectiveness of FiLMMeD, which consistently outperforms state-of-the-art baselines. Our code is available at: https://github.com/AJ-Correa/FiLMMeD/tree/main
Read the full abstract ▾
Machine Learning · Cross-listed from cs.CL
TopBench: A Benchmark for Implicit Predictive Reasoning in Tabular Question Answering
An-Yang Ji, Jun-Peng Jiang, De-Chuan Zhan et al. (+1 more)
Large
Language Models (LLMs) have advanced Table Question Answering, where most queries can be answered by extracting information or simple aggregation. However, a common class of real-world queries is implicitly predictive, requiring the inference of unobserved answers from historical patterns rather than mere
retrieval. These queries introduce two challenges: recognizing latent intent and reliable predictive
reasoning over massive tables. To assess LLMs in such Tabular questiOn answering with implicit Prediction tasks, we introduce TopBench, a
benchmark consisting of 779 samples across four sub-tasks, ranging from single-point prediction to decision making, treatment effect analysis, and complex filtering, requiring models to generate outputs spanning
reasoning text and structured tables. We evaluate diverse models under both text-based and agentic workflows. Experiments reveal that current models often struggle with intent recognition, defaulting to just lookups. Deeper analysis identifies that accurate intent disambiguation serves as the prerequisite for leading these predictive behaviors. Furthermore, elevating the upper bound of prediction precision requires the integration of more sophisticated modeling or
reasoning capabilities.
Read the full abstract ▾
All authors An-Yang Ji, Jun-Peng Jiang, De-Chuan Zhan, Han-Jia Ye
Machine Learning
A Unified Framework of Hyperbolic Graph Representation Learning Methods
Sofía Pérez Casulo, Marcelo Fiori, Bernardo Marenco et al. (+1 more)
Hyperbolic geometry has emerged as an effective latent space for representing complex networks, owing to its ability to capture hierarchical organization and heterogeneous connectivity patterns using low-dimensional
embeddings. As a result, numerous
hyperbolic graph representation learning methods have been proposed in recent years. However, their practical adoption and systematic comparison remain challenging, as implementations are fragmented and shared tools for reproducible and fair evaluation are lacking. In this work, we introduce a unified open-source framework for
hyperbolic graph representation learning that integrates several widely used
embedding methods under a common optimization interface. The novel framework enables consistent training, visualization, and evaluation of
hyperbolic embeddings, and interfaces seamlessly with standard network analysis tools. Leveraging this unified setup, we conduct an experimental study of
hyperbolic embedding methods on real-world networks, focusing on two canonical downstream tasks: link prediction and node classification. Beyond predictive accuracy, the study offers practical insights into the strengths and limitations of existing approaches, thereby facilitating informed method selection and fostering reproducible research in
hyperbolic graph representation learning.
Read the full abstract ▾
All authors Sofía Pérez Casulo, Marcelo Fiori, Bernardo Marenco, Federico Larroca
Machine Learning · Cross-listed from physics.soc-ph
Assessing the Role of Intersection Proximity in Pedestrian Crashes: Insights from Data Mining Approach
Ahmed Hossain and Xiaoduan Sun
Although intersections are the most complex parts of the roadway network, pedestrian crashes at non-intersection locations are disproportionately frequent, highlighting a serious traffic safety concern. This study investigates non-intersection crashes involving pedestrians using a crash database (2017-2021) collected from Louisiana State. As the risk of pedestrian crashes tends to vary with distance from the intersection, the research team utilized a unique framework "distance to intersection" to capture the differences in crash patterns at non-intersection locations. The study identified that around 50% of non-intersection pedestrian crashes occurred within 198 ft. of the intersection. In the next step, the collected 3,135 pedestrian crashes at non-intersection locations during the study period were subdivided into three zones: D1 zone designates crashes occurring within 150 ft. of an intersection (1,277 crashes), D2 zone designates crashes occurring within 151 ft. to 435 ft. of an intersection (1,060 crashes) and D3 zone designates crashes occurring at 435 ft. or higher from an intersection (798 crashes). To explore the complex interaction of multiple factors, an intuitive data mining technique, Association Rules Mining was used. A total of the top 60 interesting association rules (20 for each zone) were identified by the
algorithm (based on lift and support measures). In addition, a total of 124 rules were explored based on Lift Increase Criterion (LIC) measure. The findings of this research provide critical insights into pedestrian crash involvement at non-intersection locations and the variation in crash patterns according to the "distance to intersection". Based on the findings, some of the targeted problem-specific countermeasures are also recommended to address the crash patterns at non-intersection locations.
Read the full abstract ▾
Machine Learning
PROMISE-AD: Progression-aware Multi-horizon Survival Estimation for Alzheimer's Disease Progression and Dynamic Tracking
Qing Lyu, Jeremy Hudson, Mohammad Kawas et al. (+3 more)
Individualized Alzheimer's disease (AD) progression prediction requires models that use irregular visits, account for censoring, avoid diagnostic leakage, and provide calibrated horizon risks. We propose PROgression-aware MultI-horizon Survival Estimation for Alzheimer's Disease (PROMISE-AD), a leakage-safe survival framework for predicting conversion from cognitively normal (CN) to mild cognitive impairment (MCI) and from MCI to AD dementia using ADNI/TADPOLE tabular histories. PROMISE-AD converts pre-index visits into
tokens with standardized measurements, missingness masks, longitudinal changes, time-normalized slopes, visit timing, and non-diagnostic categorical attributes. A temporal
Transformer fuses global,
attention-pooled, and latest-visit representations to estimate a progression score and latent discrete-time mixture hazards. Training combines survival likelihood, horizon-specific focal risk loss, progression ranking, hazard smoothness, and mixture-balance regularization, followed by validation-set isotonic calibration for 1-, 2-, 3-, and 5-year risks. In held-out testing across three seeds, PROMISE-AD achieved an integrated Brier score (IBS) of 0.085 $\pm$ 0.012, C-index of 0.808 $\pm$ 0.015, and mean time-dependent AUC of 0.840 $\pm$ 0.081 for CN-to-MCI conversion, yielding the lowest IBS among compared methods. For MCI-to-AD conversion, PROMISE-AD achieved the highest C-index (0.894 $\pm$ 0.018) and near-ceiling 5-year discrimination (AUROC 0.997 $\pm$ 0.003; AUPRC 0.999 $\pm$ 0.001), although some baselines had lower IBS. Ablations and interpretability supported longitudinal change features, fused temporal representations, mixture hazards, cognitive and functional measures, APOE4 status, and recent conversion-proximal visits. These findings suggest that progression-aware survival modeling can provide interpretable multi-horizon AD conversion risk estimates.
Read the full abstract ▾
All authors Qing Lyu, Jeremy Hudson, Mohammad Kawas, Yuming Jiang, Chenyu You, Christopher T Whitlow
Machine Learning
Early Detection of Water Stress by Plant Electrophysiology: Machine Learning for Irrigation Management
Eduard Buss, Till Aust and Heiko Hamann
Purpose: Fast detection of plant stress is key to plant phenotyping, precision agriculture, and automated crop management. In particular, efficient irrigation management requires early identification of water stress to optimize resource use while maintaining crop performance. Direct physiological sensing offers the potential to detect stress responses before visible symptoms appear. Methods: In this study, we recorded electrophysiological signals from greenhouse-grown tomato plants subjected to water stress and developed a framework based on machine learning for online stress detection. The recorded time-series data were processed using a processing pipeline that includes statistical feature extraction and selection, automated machine learning or alternatively deep learning, and probability calibration. Results: Across multiple input time horizons, we found that a 30-minute look-back window strikes the best balance between rapid decision-making and classification performance. Using automated machine learning, the framework achieved classification accuracies of up to 92%, outperforming deep learning approaches. Sequential backward selection reduced the feature set while maintaining performance. Importantly, the framework detects transitions from healthy to stressed states in recordings that were not included in the training set. Conclusion: Overall, we provide a decision-support tool for farmers and establish a foundation for biofeedback-driven irrigation control to improve resource efficiency in (semi-)autonomous crop production systems.
Read the full abstract ▾
Machine Learning
Exponential families from a single KL identity
Marc Dymetman
Exponential families encompass the distributions central to modern machine learning – softmax, Gaussians, and Boltzmann distributions – and underlie the theory of variational inference,
entropy-regularized reinforcement learning, and RLHF. We isolate a simple identity for exponential families that expresses the KL difference $\mathrm{KL}(q \| p_{\lambda_2}) - \mathrm{KL}(q \| p_{\lambda_1})$ in terms of the log-partition function $A(\lambda)$ and the moment $\mu_q$. Remarkably, this identity together with the single fact that $\mathrm{KL} \geq 0$ (with equality iff $p = q$) suffices, by direct substitution and rearrangement, to derive a cluster of results that are classically obtained by separate, heavier arguments: a generalized three-point identity for arbitrary reference distributions, Pythagorean
theorems for I-projections and reverse I-projections, convexity of the log-partition function, identification of its Legendre dual in KL terms, the Gibbs variational principle, and the explicit optimizer in KL-regularized reward maximization, including the exponential tilting formula underlying
entropy-regularized control and RLHF. Beyond these purely algebraic consequences, standard analytic arguments recover the gradient formula for the log-partition function, the Bregman representation of within-family KL divergence, and the surjectivity of the moment map. The note is self-contained.
Read the full abstract ▾
Machine Learning
Shuffling-Aware Optimization for Private Vector Mean Estimation
Shun Takagi and Seng Pei Liew
We study $d$-dimensional unbiased mean estimation in the single-message shuffle model, where each user sends a single privatized message and the analyzer only observes the shuffled multiset of reports. While minimax-optimal mechanisms are well understood in the local differential
privacy setting, the corresponding notion of optimality after shuffling has remained largely unexplored. To address this gap, we introduce the recently proposed shuffle index and use it to formulate the post-shuffling mechanism design problem as an explicit optimization problem. We then establish a minimax lower bound on the achievable mean squared error in terms of the shuffle index, which implies that mechanisms that are optimal under LDP can become suboptimal once shuffling is applied. Finally, we construct an asymptotically minimax optimal mechanism in the high
privacy regime, which as a consequence achieves a
privacy-utility trade-off nearly identical to that of the central Gaussian mechanism.
Read the full abstract ▾
Machine Learning
MIFair: A Mutual-Information Framework for Intersectionality and Multiclass Fairness
Jeanne Monnier, Thomas George, Frédéric Guyard et al. (+2 more)
Fairness in machine learning remains challenging due to its ethical complexity, the absence of a universal definition, and the need for context-specific bias metrics. Existing methods still struggle with intersectionality, multiclass settings, and limited flexibility and generality. To address these gaps, we introduce MIFair, a unified framework for bias assessment and mitigation based on mutual information. MIFair provides a flexible metric template and an in-processing mitigation method inspired by the Prejudice Remover, defining group
fairness as statistical independence between prediction-derived variables and sensitive attributes. We further strengthen its information-theoretic foundation by establishing equivalences with widely used
fairness notions such as independence and separation. MIFair naturally supports intersectionality, complex subgroup structures, and multiclass classification and employs regularization-based training to reduce bias according to the selected metric. Its key advantage is its versatility: it consolidates diverse
fairness requirements into a single coherent framework, enabling consistent benchmarking and simplifying practical use. Experiments on real-world tabular and image
datasets show that MIFair effectively reduces bias, including previously unaddressed multi-attribute scenarios, while maintaining strong predictive performance across the evaluated settings.
Read the full abstract ▾
All authors Jeanne Monnier, Thomas George, Frédéric Guyard, Christèle Tarnec, Marios Kountouris
Machine Learning
FedHarmony: Harmonizing Heterogeneous Label Correlations in Federated Multi-Label Learning
Zhiqiang Kou, Junxiang Wu, Wenke Huang et al. (+8 more)
Federated Multi-Label Learning is a distributed paradigm where multiple clients possess heterogeneous multi-label data and perform collaborative learning under
privacy constraints without sharing raw data. However, modeling label correlations under heterogeneous distributions remains challenging. Due to client-specific label spaces and varying co-occurrence patterns, correlations learned by individual clients inevitably deviate from the global structure, a phenomenon we term label correlation drift. To address this, we propose FedHarmony, a framework that harmonizes heterogeneous label correlations across clients. It introduces consensus correlation, capturing agreement among other clients and serving as a global teacher to correct biased local estimates. During aggregation, FedHarmony evaluates each client by both data size and correlation quality, assigning weights accordingly. Moreover, we develop an accelerated optimization
algorithm for FedHarmony and theoretically establish faster convergence without sacrificing accuracy. Experiments on real-world
federated multi-label
datasets show that FedHarmony consistently outperforms state-of-the-art methods.
Read the full abstract ▾
All authors Zhiqiang Kou, Junxiang Wu, Wenke Huang, Wenwen He, Ming-Kun Xie, Changwei Wang, Yuheng Jia, Di Jiang, Yang Liu, Xin Geng, Qiang Yang
Machine Learning
Cost-Aware Learning
Clara Mohri, Amir Globerson, Haim Kaplan et al. (+2 more)
We consider the problem of Cost-Aware Learning, where sampling different components of a finite-sum objective incurs different costs. The objective is to reach a target error while minimizing the total cost. We propose Cost-Aware SGD, which uses a distribution based on gradient norms and costs to sample components. We provide a thorough analysis of this
algorithm, including cost-improvement bounds over baselines, a characterization of distribution proxy sub-optimality, and a lower bound. We apply our theoretical insights to reinforcement learning with
language models, where the computational cost of sequence-level policy gradients varies with length. We find that the advantage magnitude serves as a high-fidelity proxy for gradient norms, and use this to introduce Cost-Aware GRPO. Empirical results on 1.5B, 4B, and 8B LLMs demonstrate that this
algorithm significantly reduces the
tokens used in policy optimization while matching or exceeding baseline accuracy.
Read the full abstract ▾
All authors Clara Mohri, Amir Globerson, Haim Kaplan, Tomer Koren, Yishay Mansour
Machine Learning · Cross-listed from cs.CV
Faster 3D Gaussian Splatting Convergence via Structure-Aware Densification
Linjie Lyu, Ayush Tewari, Jianchun Chen et al. (+2 more)
3D Gaussian Splatting has emerged as a powerful scene representation for real-time novel-view synthesis. However, its standard adaptive density control relies on screen-space positional gradients, which do not distinguish between
geometric misplacement and frequency aliasing, often leading to either over-blurred high-frequency textures or inefficient over-densification. We present a structure-aware densification framework. Our key insight is that the decision to subdivide a Gaussian should be driven by an explicit comparison between its projected screen-space extent and the local structure of the texture it seeks to represent. We introduce a multi-scale frequency analysis combining structure tensors with
Laplacian scale space analysis to estimate the dominant frequency at each pixel, enabling robust supervision across varying texture scales. Based on this analysis, we define $\eta$, a per-Gaussian, per-axis frequency violation metric that indicates when a primitive may be under-resolving local texture details. Unlike methods that perform isotropic splitting (e.g., splitting each Gaussian into two smaller ones with uniform shape), our approach performs anisotropic splitting. For each axis with high $\eta$, we compute a split factor to better resolve the local frequency content. We further introduce a multiview consistency criterion that aggregates $\eta$ observations across multiple views. By performing densification early and faster, we skip the lengthy iterative densification phases required by baseline methods and achieve significantly faster convergence. Experiments on standard
benchmarks demonstrate that our method also achieves superior reconstruction quality, particularly in high-frequency regions.
Read the full abstract ▾
All authors Linjie Lyu, Ayush Tewari, Jianchun Chen, Thomas Leimkühler, Christian Theobalt
Machine Learning
Learning from Disagreement: Clinician Overrides as Implicit Preference Signals for Clinical AI in Value-Based Care
Prabhjot Singh, Abhishek Gupta, Chris Betz et al. (+4 more)
We reframe clinician overrides of clinical AI recommendations as implicit preference data - the same signal structure exploited by reinforcement learning from human feedback (RLHF), but richer: the annotator is a domain expert, the alternatives carry real consequences, and downstream outcomes are observable. We present a formal framework extending standard preference learning with three contributions: a five-category override taxonomy mapping override types to distinct model update targets; a preference formulation conditioned on patient state s, organizational context c, and clinician capability kappa, where kappa decomposes into execution capability kappa-exec and
alignment capability kappa-align; and a dual learning architecture that jointly trains a reward model and a capability model via alternating optimization, preventing a failure mode we term suppression bias-the systematic suppression of correct-but-difficult recommendations when clinician capability falls below the execution threshold. We argue that chronic disease management under outcome-based payment contracts produces override data with uniquely favorable properties-longitudinal density, concentrated decision space, outcome labels, and natural capability variation-and that training environments combining longitudinal outcome measurement with aligned financial incentives are a necessary condition for learning a reward model aligned with patient trajectory rather than with encounter economics. This framework emerged from operational work to improve clinician capability in a live value-based care deployment.
Read the full abstract ▾
All authors Prabhjot Singh, Abhishek Gupta, Chris Betz, Abe Flansburg, Brett Ives, Sudeep Lama, Jung Hoon Son
Machine Learning
Kernelized Advantage Estimation: From Nonparametric Statistics to LLM Reasoning
Shijin Gong, Kai Ye, Jin Zhu et al. (+3 more)
Recent advances in large
language models (LLMs) have increasingly relied on reinforcement learning (RL) to improve their
reasoning capabilities. Three types of approaches have been widely adopted: The first relies on a deep neural network to estimate the value function of the learning policy in order to reduce the variance of the policy gradient. However, estimating and maintaining such a value network incurs substantial computational and
memory overhead. The second avoids training a value network by approximating the value function using sample averages. However, it samples a large number of
reasoning traces per
prompt for accurate value function approximation, making it computationally expensive. The third samples only a single
reasoning trajectory per
prompt, which reduces computational cost but suffers from poor sample efficiency. This paper focuses on a practical, resource-constrained setting in which only a small number of
reasoning traces can be sampled per
prompt, while low-variance gradient estimation remains essential for high-quality policy learning. To address this challenge, we bring classical nonparametric statistical methods, which are both computationally and statistically efficient, to LLM
reasoning. We employ kernel smoothing as a concrete example for value function estimation and the subsequent policy optimization. Numerical and theoretical results demonstrate that our proposal achieves accurate value and gradient estimation, leading to improved policy optimization.
Read the full abstract ▾
All authors Shijin Gong, Kai Ye, Jin Zhu, Xinyu Zhang, Hongyi Zhou, Chengchun Shi
Machine Learning
Latent-GRPO: Group Relative Policy Optimization for Latent Reasoning
Jingcheng Deng, Zihao Wei, Liang Pang et al. (+4 more)
Latent
reasoning offers a more efficient alternative to explicit
reasoning by compressing intermediate
reasoning into continuous representations and substantially shortening
reasoning chains. However, existing latent
reasoning methods mainly focus on supervised learning, and reinforcement learning in latent space remains highly unstable. We study this problem through the lens of Group Relative Policy Optimization (GRPO), and show that directly adapting GRPO to latent
reasoning is fundamentally non-trivial: latent
reasoning changes both the probability density and the sampling mechanism, causing three coupled bottlenecks: absence of intrinsic latent
manifolds, where unconstrained exploration pushes rollouts off the valid latent
manifold; exploration-optimization misalignment, where trajectory-level rewards can induce incorrect
token-level updates; and latent mixture non-closure, where jointly reinforcing multiple correct latent paths can produce an invalid averaged state. To address them, we propose Latent-GRPO, which combines invalid-sample advantage masking, one-sided noise sampling, and optimal correct-path first-
token selection. Across four low-difficulty
benchmarks (e.g., GSM8K-Aug) and four high-difficulty
benchmarks (e.g., AIME), Latent-GRPO improves over its latent initialization by 7.86 Pass@1 points on low-difficulty tasks and surpasses explicit GRPO by 4.27 points on high-difficulty tasks while using 3–4$\times$ shorter
reasoning chains. It also achieves stronger pass@$k$ performance under Gumbel sampling. These results establish Latent-GRPO as an effective approach for stable and efficient latent
reasoning.
Read the full abstract ▾
All authors Jingcheng Deng, Zihao Wei, Liang Pang, Junhong Wu, Shicheng Xu, Zenghao Duan, Huawei Shen
Machine Learning
Dynamic Scaled Gradient Descent for Stable Fine-Tuning for Classifications
Nghia Bui and Lijing Wang
Fine-tuning pretrained models has become a standard approach to adapting pretrained knowledge to improve the accuracy on new sparse, imbalance
datasets. However, issues arise when optimization falls into a collapsed state, where the model gets stuck, leading to degraded performance and unstable training. One possible reason for this is the cancellation of gradients across training examples. To address this problem, we propose a novel
algorithm, dynamic scaled gradient descent (\mName), that directly modifies the gradients returned by training examples, specifically, scaling down the gradients of correctly classified examples using a dynamic scaler. This strategy offers both theoretical and empirical advantages in improving training stability. Experiments on a variety of
benchmark datasets, spanning multiple tasks and large pretrained models, demonstrate that our method consistently reduces performance variance and surpasses the accuracy of existing approaches.
Read the full abstract ▾
Machine Learning
ITS-Mina: A Harris Hawks Optimization-Based All-MLP Framework with Iterative Refinement and External Attention for Multivariate Time Series Forecasting
Pourya Zamanvaziri, Amirhossein Sadr, Aida Pakniyat et al. (+1 more)
Multivariate time series
forecasting plays a pivotal role in numerous real-world applications, including financial analysis, energy management, and traffic planning. While
Transformer-based architectures have gained popularity for this task, recent studies reveal that simpler MLP-based models can achieve competitive or superior performance with significantly reduced computational cost. In this paper, we propose ITS-Mina, a novel all-MLP framework for multivariate time series
forecasting that integrates three key innovations: (1) an iterative refinement mechanism that progressively enhances temporal representations by repeatedly applying a shared-parameter residual mixer stack, effectively deepening the model's computational capacity without multiplying the number of distinct parameters; (2) an external
attention module that replaces traditional self-
attention with learnable
memory units, capturing cross-sample global dependencies at linear computational complexity; and (3) a Harris Hawks Optimization (HHO)
algorithm for automatic dropout rate tuning, enabling adaptive regularization tailored to each
dataset. Extensive experiments on six widely-used
benchmark datasets demonstrate that ITS-Mina achieves state-of-the-art or highly competitive performance compared to eleven baseline models across multiple
forecasting horizons.
Read the full abstract ▾
All authors Pourya Zamanvaziri, Amirhossein Sadr, Aida Pakniyat, Dara Rahmati
Machine Learning · Cross-listed from cs.AI
D3-Gym: Constructing Real-World Verifiable Environments for Data-Driven Discovery
Hanane Nour Moussa, Yifei Li, Zhuoyang Li et al. (+7 more)
Despite recent progress in
language models and
agents for scientific data-driven discovery, advancing their capabilities is held back by the absence of verifiable environments representing real-world scientific tasks. To fill this gap, we introduce D3-Gym, the first automatically constructed
dataset with verifiable environments for scientific Data-Driven Discovery. D3-Gym comprises 565 tasks from 239 real scientific repositories across four disciplines, each with a natural
language instruction, an executable environment with pre-installed dependencies,
dataset previews, a reference solution, and an automatically synthesized evaluation script. Our evaluation scripts achieve 87.5% agreement with human-annotated gold standards and strong
alignment in domain-specific evaluation logic. Training on trajectories sampled from D3-Gym yields consistent gains across Qwen3 models on ScienceAgentBench, boosting Qwen3-32B by 7.8 absolute points and shrinking the gap with strong proprietary models. We further illustrate, through case studies, how D3-Gym environments can serve as a testbed for studying agentic optimization loops such as Autoresearch on real scientific workflows. We open-source D3-Gym, its creation workflow, sampled trajectories, and training scripts at https://github.com/OSU-NLP-Group/D3-Gym.
Read the full abstract ▾
All authors Hanane Nour Moussa, Yifei Li, Zhuoyang Li, Yankai Yang, Cheng Tang, Tianshu Zhang, Nesreen K. Ahmed, Ali Payani, Ziru Chen, Huan Sun
Machine Learning
Differentiable latent structure discovery for interpretable forecasting in clinical time series
Ivan Lerner, Jean Feydy, Alexandre Kalimouttou et al. (+2 more)
Background: We introduce StructGP, a continuous-time multi-task Gaussian process that couples process convolutions with differentiable structure learning to uncover a sparse, ordered directed acyclic graph (DAG) of inter-variable dependencies while preserving principled uncertainty. We further propose LP-StructGP, which augments StructGP with latent pathways-shared, temporally shifted trajectories inferred via subject-specific coupling filters and a softmax gating mechanism-to capture cross-patient progression patterns. Both models are trained under sparsity and acyclicity constraints using scalable low-rank updates using likelihood-based objectives. Results: In simulations, graph recovery improved with cohort size, with the median Structural Hamming Distance reaching zero at the largest cohort size, while pathway assignments showed high Adjusted Rand Index. Our analysis establishes that the ordered StructGP graph is identifiable from the population marginal likelihood. On a MIMIC-IV septic shock cohort (n=1,008; norepinephrine, creatinine, mean blood pressure), StructGP improves short-horizon (6 h)
forecasting over independent-task baselines (average RMSE 0.68 [95% CI: 0.63-0.74] vs. 0.88 [0.83-0.94]) and, with 15 additional inputs, markedly outperforms unstructured kernels (0.63 [0.58-0.69] vs. 3.02 [2.85-3.18]) with superior calibration (coverage 0.96 vs. 0.84). For long horizons (up to 6 days), LP-StructGP further reduces error for creatinine (RMSE 0.95 [0.88-1.03] vs. 1.17 [1.08-1.25]) and improves overall coverage (0.93 [0.93-0.94] vs. 0.91 [0.91-0.92]). On the PhysioNet Challenge, StructGP attains competitive accuracy (MAE 3.72e-2) relative to a strong published graph neural model. Conclusion: These results show that structured process convolutions with latent pathways deliver interpretable, scalable, and well-calibrated
forecasting for irregular clinical time series.
Read the full abstract ▾
All authors Ivan Lerner, Jean Feydy, Alexandre Kalimouttou, Anita Burgun, Francis Bach
Machine Learning · Cross-listed from eess.IV
Diffusion-OAMP for Joint Image Compression and Wireless Transmission
Wentao Hou, Yimin Bai, Zelei Luo et al. (+2 more)
Joint image compression and wireless transmission remain relatively underexplored compared to generic image restoration, despite its importance in practical communication systems. We formulate this problem under an equivalent linear model, and propose
Diffusion-OAMP, a training-free reconstruction framework that embeds a pre-trained
diffusion model into the OAMP
algorithm. In
Diffusion-OAMP, the OAMP linear
estimator produces pseudo-AWGN observations, while the
diffusion model serves as a nonlinear
estimator under an SNR-matching rule. This framework offers a way to incorporate multiple generative priors into OAMP. Experiments with varying compression ratios and noise levels show that
Diffusion-OAMP performs favorably against classic methods in the evaluated settings.
Read the full abstract ▾
All authors Wentao Hou, Yimin Bai, Zelei Luo, Jiadong Hong, Lei Liu
Machine Learning · Cross-listed from cs.NE
Attractor FCM
Alexis Kafantaris
In this paper an attractor FCM is created, tested, and analyzed. This FCM is neither a hebbian based nor agentic, nor a hybrid; it rather is a gradient descent based, physics constrained, Jacobian version of an FCM. Moreover, this model has several quirks; it uses residual
memory, back propagation through time, and a fixed point anchor that is recursively implemented to update its weights. The residuals update the recursive part without losing the system
memory. The model's anchor enables it to converge in a fixed point for which back propagation through time unrolls it and ensures that the error minimization is for an accurate gradient. Furthermore, a new learning
algorithm is utilized. The Newton's method finds the system's fixed point attractor and then gradient descend is adaptively changing the landscape; an adaptive term is used to directly manipulate the weights through the attractor dynamics. As the adaptive term changes, the descent through the landscape is constantly adjusting according to sigmoid saturation, and that prevents premature convergence to a local minimum. Lastly, the updates are filtered by causal mask that informs the network about the physics, respecting the initial expert based opinions, for which model reduces the error to the target in an efficient way.
Read the full abstract ▾
Machine Learning
Calibrating Attribution Proxies for Reward Allocation in Participatory Weather Sensing
Mark C. Ballandies, Michael T. C. Chiu and Claudio J. Tessone
Large-scale IoT weather sensing networks require incentive mechanisms to sustain participation, yet determining how much value individual data contributions bring to the network remains an open problem. Existing approaches address data quality but not data valuation; in operational meteorology, adjoint-based methods derive value from the forecast model itself but require full data assimilation infrastructure. We propose to utilise differentiable AI weather models to fill this gap and characterise gradient-based attribution on gridded GFS analysis inputs as a candidate value signal, evaluating fidelity, calibration, cost, and gaming vulnerability across more than 400 configurations. Attribution captures near-optimal sensor placement utility with monotonically faithful payments, but can be inflated by
adversarial inputs, with detection requiring external baseline data. These findings establish gradient attribution as a computationally validated signal for model-informed reward allocation in participatory weather sensing.
Read the full abstract ▾
Machine Learning
Beyond the Baseband: Adaptive Multi-Band Encoding for Full-Spectrum Bioacoustics Classification
Eklavya Sarkar, Marius Miron, David Robinson et al. (+6 more)
Animals hear and vocalize across frequency ranges that differ substantially from humans, often extending into the ultrasonic domain. Yet most computational bioacoustics systems rely on audio models pre-trained at 16 kHz, restricting their usable bandwidth to the 0-8 kHz baseband and discarding higher-frequency information present in many bioacoustic recordings. We investigate a multi-band encoding framework that decomposes the full spectrum of animal calls into band features and fuses them into a unified representation. Similarity analyses on models show that certain encoders produce decorrelated band
embeddings that improve class separation after fusion. Classification experiments on three bioacoustic
datasets using eight pre-trained models and five fusion strategies show that fused representations consistently outperform the baseband and time-expansion baselines on two
datasets, showing the potential of multi-band methods for full-spectrum encoding of animal calls.
Read the full abstract ▾
All authors Eklavya Sarkar, Marius Miron, David Robinson, Gagan Narula, Milad Alizadeh, Ellen Gilsenan-McMahon, Emmanuel Chemla, Olivier Pietquin, Matthieu Geist