cs.LG · 2026-05-30 · No. 12

Machine Learning, 2026-05-30.

64 new papers in cs.LG. Titles, authors, abstracts. Links to arXiv. Want this in your inbox every morning? Subscribe →

01 — The papers

64 entries
  1. 01

    Efficient Test-Time Finetuning of LLMs via Convex Reconstruction and Gradient Caching

    Alaa Khamis, Alaa Maalouf

    cs.LG

    Test-time finetuning (TTFT) is a rapidly evolving paradigm that adapts a language model to each prompt by retrieving related sequences, updating the model on them, and then evaluating the prompt. However, TTFT is only practical if it is fast: selection and finetuning both happen per query, making each a direct bottleneck. Existing methods trade speed for quality: fast retrieval is often redundant, while stronger diversity-aware selection adds...

    arxiv.org/abs/2605.30337 · PDF

  2. 02

    Fairness-Aware Federated Learning with Trajectory Shapley Value

    Daniel Kuznetsov, Ziqi Wang

    cs.LG

    Federated learning is an emerging distributed paradigm that addresses the challenges posed by heterogeneous, privacy-sensitive data. It enables multiple clients to train a model collaboratively by aggregating their local updates at a server. However, conventional aggregation schemes typically use fixed weights that fail to reflect unequal and time-varying client contributions, leading to biased and unstable learning. To improve fairness and...

    arxiv.org/abs/2605.30336 · PDF

  3. 03

    When, why, and how do diffusion posterior samplers fail? A finite-sample lens

    Benjamin A. Burns, Sara Fridovich-Keil

    cs.LG

    Diffusion models have excellent capacity to model complex distributions of natural data, which has made them a popular and effective choice for posterior sampling in imaging inverse problems. Existing methods can incorporate any measurement model at inference time but must use an inexact approximation for the likelihood at intermediate timesteps for computational tractability. Although these approximations can often work well empirically,...

    arxiv.org/abs/2605.30330 · PDF

  4. 04

    SoundnessBench: Can Your AI Scientist Really Tell Good Research Ideas from Bad Ones?

    Sy-Tuyen Ho, Minghui Liu, Huy Nghiem, Furong Huang

    cs.LG

    Autonomous AI research agents aim to accelerate scientific discovery by automating the research pipeline, from hypothesis generation to peer review. However, existing benchmarks rarely test a fundamental bottleneck: whether Large Language Models can judge the methodological viability of a research idea before expending time and computational resources. We introduce SoundnessBench, a curated benchmark of 1,099 machine-learning research...

    arxiv.org/abs/2605.30329 · PDF

  5. 05

    Reasoning with Sampling: Cutting at Decision Points

    Felix Zhou, Anay Mehrotra, Quanquan C. Liu

    cs.LG · cs.AI · cs.CL · math.ST · stat.ML

    Frontier reasoning models are produced by posttraining base language models with reinforcement learning. Recent work has challenged this by showing that sampling from a sharpened version of the base model's distribution, a so-called power distribution, elicits comparable reasoning without additional training, curated datasets, or verifiers. However, making this method practical requires efficiently sampling from the power distribution. A...

    arxiv.org/abs/2605.30327 · PDF

  6. 06

    In-Context Reward Adaptation for Robust Preference Modeling

    Zhenyu Sun, Zheng Xu, Ermin Wei

    cs.LG · cs.AI

    Reinforcement Learning from Human Feedback (RLHF) typically relies on static reward models to align Large Language Models with human preferences. However, human values are inherently diverse and heterogeneous, and a single reward model often lacks the robustness required to generalize to unseen preference domains. While existing multi-reward frameworks attempt to address this, they are often restricted to a fixed set of known domains and fail...

    arxiv.org/abs/2605.30323 · PDF

  7. 07

    Gram: Assessing sabotage propensities via automated alignment auditing

    David Lindner, Victoria Krakovna, Sebastian Farquhar

    cs.LG · cs.AI

    We introduce Gram, an automated alignment auditing framework to assess the propensity of AI agents to engage in sabotage. We evaluate Gemini models across 17 simulated agentic deployment scenarios that incentivize sabotage. We find Gemini models misbehave in about 2-3% of our simulated trajectories. Many of these cases are explained by "overeagerness" in Gemini models resulting in both excessive role-playing and goal-seeking behavior. In...

    arxiv.org/abs/2605.30322 · PDF

  8. 08

    Self-Trained Verification for Training- and Test-Time Self-Improvement

    Chen Henry Wu, Aditi Raghunathan

    cs.LG · cs.AI · cs.CL

    Self-improvement at scale has been a longstanding goal for reasoning models, and there are two natural places to do it: at test time, through verification-refinement (V-R) loops; and at training time, through self-training methods. Both are gated by the same bottleneck: the verifier. V-R loops stall when verifier scores inflate while accuracy stagnates, and when feedback is too generic to act on; self-training fails similarly when bad...

    arxiv.org/abs/2605.30290 · PDF

  9. 09

    Statistical Embeddings for Similarity, Retrieval, and Interpretable Alignment of Numeric Tabular Datasets

    M. Ross Kunz, John Merickel, Keith Wilson

    cs.LG · stat.AP · stat.ML

    Numeric tabular datasets are the dominant data format in scientific practice, yet large language models lack native mechanisms for representing numeric datasets in a meaningful way across heterogeneous feature spaces. Existing approaches either target predictive modeling over individual datasets, which requires a shared set of variable definitions, or lack mechanisms for interpretable cross-dataset alignment. The proposed methodology...

    arxiv.org/abs/2605.30289 · PDF

  10. 10

    Neural Operator-Based Surrogate Model for CFD:Helical Coil Steam Generator in Small Modular Reactor

    Minseo Lee, Seongmin Oh, Chaehyeon Song, Bumjin Cho, Shilaj Baral, Sangam Khanal, Minseop Song, Joongoo Jeon

    cs.LG · physics.flu-dyn

    Real-time thermal-hydraulic simulation is essential for digital twin (DT) technology that supports the safe and efficient operation of small modular reactors (SMRs). Computational fluid dynamics (CFD) provides high-fidelity flow analysis, but its computational cost prevents direct use in DT applications. AI-based surrogate modeling has been actively investigated to address this limitation, yet neural operator--based surrogates for CFD-level...

    arxiv.org/abs/2605.30277 · PDF

  11. 11

    Digitally enriching a screening population for pancreatic cancer using routine blood-based measures and clinical histories

    Chris Varghese, Leo Y. Li-Han, Richa Bisht, Ellen Larson, Frank Lee, Ryan M. Carr, Tanios S. Bekaii-Saab, Shounak...

    cs.LG · q-bio.QM

    Earlier detection of pancreatic cancer is key to enabling wider access to curative treatment and reducing cancer deaths; however, screening is presently not viable. Latent indicators of pathology are evident in an individual's disease and blood test trajectories and may predict the development of pancreatic cancer. Longitudinal sequences of coded diagnoses and blood test values accrued by patients throughout their clinical interactions were...

    arxiv.org/abs/2605.30275 · PDF

  12. 12

    OOD-GraphLLM: Graph Large Language Model for Out-of-Distribution Generalized Drug Synergy Prediction

    Xin Wang, Linxin Xiao, Yang Yao, Wenwu Zhu

    cs.LG · cs.MM

    Drug synergy prediction (DSP) aims to identify efficacious drug combinations under various cellular contexts with different targets. However, the continual emergence of novel compounds results in variations in molecular scaffolds and sizes, causing drug synergy data to exhibit out-of-distribution (O.O.D.) shifts with respect to topological structure. Existing works rely on in-distribution (I.D.) assumption, failing to handle the O.O.D....

    arxiv.org/abs/2605.30247 · PDF

  13. 13

    How's it going? Reinforcement learning in language models recruits a functional welfare axis

    Andy Q Han, David J. Chalmers, Pavel Izmailov

    cs.LG · cs.CL

    How does reinforcement learning shape a language model's internal representations? We present evidence that RL recruits a pre-existing representation of functional welfare: an estimate of how well or badly the system is doing, relative to its goals. We train several language models in a novel, semantically neutral maze environment. We then extract concept vectors for rewarded and punished trajectories, and evaluate those vectors in settings...

    arxiv.org/abs/2605.30232 · PDF

  14. 14

    Anti Mode-Collapse in Mean-Field Transformer via Auxiliary Variables

    Masaaki Imaizumi, Masanori Koyama, Noboru Isobe, Kohei Hayashi

    cs.LG

    We use a mean-field-based transformer model to theoretically investigate how auxiliary variables, such as positional encoding, prevent mode collapse of self-attention mechanisms. The use of mean-field transformers to analyze the properties of self-attention mechanisms has garnered significant attention in recent years due to their ability to comprehensively analyze token interactions. However, analysis of this simple model suggests that mode...

    arxiv.org/abs/2605.30229 · PDF

  15. 15

    ExDBSCAN: Explaining DBSCAN with Counterfactual Reasoning -- Additional Material

    Pernille Matthews, Lena Krieger, Tommaso Amico, Artur Zimek, Thomas Seidl, Ira Assent

    cs.LG

    Clustering is an unsupervised technique for grouping data points by similarity. While explainability methods exist for supervised machine learning, they are not directly applicable to clustering, making it challenging to understand cluster assignments. This interpretability gap is particularly evident in the popular density-based method DBSCAN, which assigns points as inliers (cluster members in dense regions) or outliers (noise points in...

    arxiv.org/abs/2605.30225 · PDF

  16. 16

    TriSearch: Learning to Optimize Triangulations via Bistellar Flips

    Yiran Wang, Guido Montúfar

    cs.LG

    We introduce TriSearch, a reinforcement learning framework for optimizing objectives over triangulations of a polytope via bistellar flips. The key idea is a circuit-supported subtriangulation action representation: feasible flips are encoded by their supporting circuit and realized local subtriangulation, enabling a learned policy to rank them using local geometric and combinatorial features. This yields a dimension-agnostic interface and...

    arxiv.org/abs/2605.30220 · PDF

  17. 17

    MarginGate: Sparse Margin-Triggered Verification for Batch-Invariant LLM Inference

    Kexin Chu, Yang Zhou, Wei Zhang

    cs.LG · cs.PF

    Temperature-zero BF16 LLM inference is often treated as reproducible, yet the same request can emit different tokens when decoded alone or inside a larger batch. Existing fixes use batch-invariant operators or LLM-42's per-token verification, incurring cost even when most steps are stable. We ask whether verification can be applied exclusively to flipped tokens. Across five models, batch-induced token flips are sparse on the flip-rate...

    arxiv.org/abs/2605.30218 · PDF

  18. 18

    Faithful Embeddings of Irregular and Asynchronous Data for Online Log-NCDEs

    Benjamin Walker, Alexandre Bloch, Lingyi Yang, Sam Morley, Terry Lyons

    cs.LG

    Continuous-time models are a natural choice for irregular and asynchronous data. A central design choice is how to embed discrete observations into continuous time. Interpolation- and imputation-based embeddings reconstruct a continuous observation path, making the model sensitive to the choice of reconstruction. We show that this reconstruction step is unnecessary; under mild conditions, compact-set universality on the model input space...

    arxiv.org/abs/2605.30213 · PDF

  19. 19

    HPO: Hysteretic Policy Optimization for Stable and Efficient Training under Sparse-Reward Regime

    Mohamed Sana, Nicola Piovesan, Antonio De Domenico, Fadhel Ayed, Haozhe Zhang

    cs.LG · cs.AI

    We investigate a narrow but common failure mode of GRPO-style reinforcement learning in the context of sparse verifiable rewards: early updates contain more responses with negative advantages than those with positive advantages, while response-level length normalization ties the magnitude of the update to the length of the output. We propose Hysteretic Policy Optimization (HPO), a minimal modification of GRPO that reduces the weight of...

    arxiv.org/abs/2605.30201 · PDF

  20. 20

    Active Continual Learning with Metaplastic Binary Bayesian Neural Networks

    Kellian Cottart, Théo Ballet, Djohan Bonnet, Damien Querlioz

    cs.LG

    Always-on edge systems must keep learning as conditions change under tight compute budgets and must detect unreliable predictions. Bayesian binary neural networks are attractive in this setting, but mean-field Bernoulli posteriors can saturate on long non-stationary streams, wiping out epistemic uncertainty and freezing plasticity. We propose BiMU, derived from a bounded-memory variational objective that balances stability, plasticity, and...

    arxiv.org/abs/2605.30198 · PDF

  21. 21

    Mean-Field Diffuser: Scaling Offline MARL to Thousands of Agents

    Wenhao Li, Xiangfeng Wang, Bo Jin

    cs.LG

    Diffusion-based planning has achieved strong results in single-agent offline reinforcement learning, yet scaling to many-agent systems remains intractable due to the curse of dimensionality in the joint trajectory space. We introduce MF-Diffuser, a framework that lifts trajectory planning to the Wasserstein space of trajectory distributions, where the propagation of chaos ensures a small representative subset of agents captures the full...

    arxiv.org/abs/2605.30190 · PDF

  22. 22

    CalArena: A Large-Scale Post-Hoc Calibration Benchmark

    Eugène Berta, David Holzmüller, Francis Bach, Michael I. Jordan

    cs.LG · cs.AI · stat.ML

    Reliable probability estimates are critical in many machine learning applications, yet modern classifiers are often poorly calibrated. Post-hoc calibration provides a simple and widely used solution, but the large number of proposed methods, combined with small-scale and inconsistent evaluations, makes it difficult to determine which approaches are truly effective in practice. We introduce a large-scale, standardized benchmark for post-hoc...

    arxiv.org/abs/2605.30188 · PDF

  23. 23

    Can AI Weather Models Predict Beyond Two Weeks? A Quantitative Benchmark and Analysis of Long Rollouts

    Fanny Lehmann, Firat Ozdemir, Yun Cheng, Torsten Hoefler, Sebastian Schemm, Benedikt Soja, Siddhartha Mishra

    cs.LG · physics.ao-ph

    While AI weather models excel at short-to-medium range forecasts (up to 15 days), they frequently suffer from ill-defined "instabilities" when rolled out over longer horizons. This work addresses the lack of a formal taxonomy by categorizing these failures into three distinct regimes: blow-up, drift, and loss of seasonality, through year-long rollouts of nine state-of-the-art AI weather models. Our analysis reveals that stability hinges on...

    arxiv.org/abs/2605.30184 · PDF

  24. 24

    iLoRA: Bayesian Low-Rank Adaptation with Latent Interaction Graphs for Microbiome Diagnosis

    Yang Song, Yixuan Zhang, Lingfa Meng, Tongyuan Hu, Haizhou Shi, Hao Wang, Samir Bhatt, Hengguan Huang

    cs.LG · cs.AI

    Parameter-efficient adaptation has made LLMs practical for domain prediction, but standard LoRA still relies on a static low-rank update and does not expose the latent interactions that often drive scientific labels. We introduce iLoRA. To our knowledge, it is the first Bayesian graph-conditioned LoRA framework. It infers a latent interaction graph from the input and uses it to generate input-conditioned LoRA updates. As a result, iLoRA...

    arxiv.org/abs/2605.30179 · PDF

  25. 25

    On Distributional Reinforcement Learning in Chaotic Dynamical Systems

    James Rudd-Jones, Mirco Musolesi, María Pérez-Ortiz

    cs.LG · cs.AI

    Chaotic dynamical systems pose a fundamental challenge for Reinforcement Learning (RL): exponential sensitivity to initial conditions induces high-variance bootstrap targets and poorly conditioned gradient updates. Chaotic dynamics arise across scientific and engineering domains, from fluid flows and climate systems to multi-agent systems, where reliable learning is highly desirable. Standard RL methods optimise expected returns through...

    arxiv.org/abs/2605.30160 · PDF

  26. 26

    RL2ML: Finite-Rollout Surrogate Objectives from Reinforcement Learning to Maximum Likelihood

    Yifu Zheng

    cs.LG

    Correctness-based Reinforcement Learning with Verifiable Rewards (RLVR) trains language models from binary feedback on sampled outputs, but the objective optimized in expectation and the stochastic update geometry induced by finite rollout groups are often conflated. This paper develops RL2ML, a family of finite-rollout surrogate objectives with a closed-form, exactly unbiased gradient estimator. The family continuously connects standard...

    arxiv.org/abs/2605.30154 · PDF

  27. 27

    Overcoming Forgetting in LLM Fine-Tuning with Evolution Strategies

    Kajetan Schweighofer, Conor F. Hayes, Roberto Dailey, Risto Miikkulainen, Xin Qiu

    cs.LG · cs.AI

    Evolution Strategies (ES) has recently emerged as a competitive alternative to reinforcement learning (RL) for large language model (LLM) fine-tuning, offering advantages through simplicity, scalability, and inference-only training. However, recent work suggests that ES fine-tuning on new tasks may induce forgetting of prior tasks. First, this paper shows that prior task forgetting (1) is better characterized as performance drift rather than...

    arxiv.org/abs/2605.30148 · PDF

  28. 28

    DAMEL: Dual-Axis Multi-Expert Learning for Class-Imbalanced Learning

    Hyuck Lee, Taemin Park, Heeyoung Kim

    cs.LG · cs.AI

    Various algorithms have been proposed to address the challenges posed by class-imbalanced learning from real-world data with long-tailed distributions. While these algorithms reduce prediction bias through rebalancing techniques, they often introduce increased prediction variance as a trade-off. Several multi-expert learning algorithms aim to address this variance but involve complex procedures. We propose a new multi-expert learning...

    arxiv.org/abs/2605.30135 · PDF

  29. 29

    Learning to Extrapolate to New Tasks: A Relational Approach to Task Extrapolation

    Adam Ousherovitch, Yixin Wang

    cs.LG · stat.ML

    Modern learning systems excel at interpolation but struggle to generalize to unseen tasks outside the training distribution's support. This failure occurs even in simple settings, such as handling task parameters beyond the training range, and persists despite advances in foundation models. To this end, we develop the Relational Task Extrapolator (RTE), an algorithm designed to enable systematic extrapolation to novel tasks. The key...

    arxiv.org/abs/2605.30132 · PDF

  30. 30

    Beyond MSE: Improving Precipitation Nowcasting with Multi-Quantile Regression

    Gijs van Nieuwkoop, Siamak Mehrkanoon

    cs.LG · cs.AI

    Deep-learning precipitation nowcasting models are often optimized using pointwise losses such as mean squared error or mean absolute error, which can lead to overly smooth forecasts and poor representation of heavy rainfall. This study investigates whether the predictive performance of an established deterministic nowcasting architecture can be improved by reformulating training as a multi-quantile regression problem. Using SmaAt-UNet as a...

    arxiv.org/abs/2605.30122 · PDF

  31. 31

    Evolving Features vs Evolving Entire Trees with GP for Interpretable Survival Analysis

    Thalea Schlender, Peter A. N. Bosman, Tanja Alderliesten

    cs.LG · cs.AI · cs.NE

    Survival analysis concerns the task of predicting the time until an event occurs. Often used in the medical field, survival analysis deals with incomplete (i.e., censored) data, for instance, from patients who did not experience the event during the duration of the study. For practical use, both accuracy and interpretability are important. Survival trees are easy-to-follow survival models that split the patient cohort recursively into...

    arxiv.org/abs/2605.30119 · PDF

  32. 32

    Striding Across Reynolds Numbers: Representation Geometry in Neural PDE Generalisation

    Jianing Shi

    cs.LG

    Cross-Reynolds generalisation in neural PDE solvers remains poorly characterised. On the canonical forced 2D Navier-Stokes benchmark, a trained Fourier Neural Operator reaches 46.68% relative L2 error under a 10x Reynolds-number shift, yet zero-forward-model retrieval baselines already improve to 41-42%. This suggests representation geometry as a major organising variable among the tested methods. We test this hypothesis through ConvAE-Relay,...

    arxiv.org/abs/2605.30112 · PDF

  33. 33

    Convergence Theory for Iterative LLM-Based Neural Architecture Search: A Parametric Cross-Entropy Framework with Closed-Form Proxy Reliability

    Santosh Premi Adhikari, Radu Timofte, Dmitry Ignatov

    cs.LG

    Large language models (LLMs) are increasingly used as generators in iterative neural architecture search (NAS), yet no formal convergence theory exists for this class of algorithms. We model iterative LLM-NAS as a parametric Cross-Entropy (CE) method over executable programs and prove six results: (1) iterative LLM fine-tuning on elite architectures is equivalent to the CE update restricted to the LLM parametric family; (2) expected...

    arxiv.org/abs/2605.30103 · PDF

  34. 34

    Chess-World-Model: A 10M-Game Benchmark for Exact State Tracking from Chess Move Sequences

    Benjamin Walker, Terry Lyons

    cs.LG

    World models require state tracking, which is the ability to maintain a correct latent state across action sequences. Existing benchmarks are often synthetic or language-based, limiting their value as tests of structured state updates in realistic domains. We introduce Chess-World-Model, a large-scale state-tracking benchmark built from 10 million real chess games, where models predict the exact board state reached after a sequence of legal...

    arxiv.org/abs/2605.30100 · PDF

  35. 35

    Distributionally Robust Set Representation Learning Under Inference-Time Element Corruption

    Yankai Chen, Hanrong Zhang, Bowei He, Philip S. Yu, Xue, Liu

    cs.LG

    Standard Set Representation Learning methods typically excel on curated data but often overlook the challenge of inference-time element corruption. This refers to scenarios where deployed models encounter element-level degradations, such as outliers or missing components, that may distort set representation and degrade performance. We propose SW-DRSO, a distributionally robust optimization framework tailored for sets. Rather than minimizing...

    arxiv.org/abs/2605.30089 · PDF

  36. 36

    Q-ANCHOR: Federated Quantum Learning with ZNE-guided Correction

    Hoang M. Ngo, Quan Nguyen, Wanli Xing, My T. Thai

    cs.LG · cs.DC

    Quantum Federated Learning (QFL) offers a promising framework to train quantum models across distributed clients while keeping data strictly local. Due to its simplicity and low communication overhead, Federated Averaging (FedAvg) is the standard aggregation choice in QFL literature. However, deploying QFL on practical hardware exposes a severe double-drift phenomenon: the global model is simultaneously derailed by client drift from non-IID...

    arxiv.org/abs/2605.30075 · PDF

  37. 37

    A Predictive Law for On-Policy Self-Distillation From World Feedback

    Tommy He, Jerome Sieber, Matteo Saponati

    cs.LG · cs.AI

    Moving beyond simple scalar rewards toward richer world feedback is a natural path to more scalable RL post-training. On-policy self-distillation (OPSD) is a promising recent approach that uses arbitrary feedback as learning signal, yet its reliability compared to established methods, such as GRPO, remains unclear. We identify a strikingly consistent linear correlation between the initial student-self-teacher performance gap and the final...

    arxiv.org/abs/2605.30070 · PDF

  38. 38

    Ridge Regression from Poisson Resetting: A Renewal Perspective on Spectral Regularization

    Petar Jolakoski

    cs.LG · cond-mat.stat-mech · stat.ML

    We connect stochastic resetting from non-equilibrium statistical physics with ridge regularization in statistical learning. For linear gradient flow, resetting to the origin at rate $r$ produces stationary mean $(X^\top X+rI)^{-1}X^\top y$, exactly the ridge estimator with penalty $λ=r$. This uses the known Laplace-transform relationship between ridge regression and exponential-time averaging of gradient flow, with the exponential time now...

    arxiv.org/abs/2605.30059 · PDF

  39. 39

    Masked Diffusion Modeling for Anomaly Detection

    Lixing Zhang, Yuchen Liang, Liyan Xie

    cs.LG · cs.AI

    Anomaly detection aims to identify samples that deviate from the nominal data distribution and is central to many safety-critical applications. However, developing effective anomaly detection methods for categorical, mixed-type, and discrete sequence data remains challenging and relatively underexplored. Masked diffusion models provide a natural way to model such data by learning to recover masked values from the remaining visible context. In...

    arxiv.org/abs/2605.30046 · PDF

  40. 40

    Alignment-Guided Score Matching for Text-to-Image Alignment in Diffusion Models

    Jaa-Yeon Lee, Yeobin Hong, Taesung Kwon, Jong Chul Ye

    cs.LG · cs.AI · cs.CV

    Diffusion models generate highly realistic images but often struggle with precise text-image alignment. While recent post-training methods improve alignment using external rewards or human preference signals, their performance heavily depends on reward quality and does not directly address alignment within the diffusion process itself. Recent reward-free approaches such as SoftREPA demonstrate that optimizing soft text tokens via contrastive...

    arxiv.org/abs/2605.30038 · PDF

  41. 41

    Test Time Training for Supervised Causal Learning

    Zizhen Deng, Jiaru Zhang, Rui Ding, Huang Bojun, Jinzhuo Wang, Qiang Fu, Shi Han, Dongmei Zhang

    cs.LG · cs.AI

    Supervised Causal Learning (SCL) has shown promise in causal discovery by framing it as a supervised learning problem. However, it suffers from significant out-of-distribution generalization challenges. We reveal three limitations of previous SCL practices: a significant performance gap between synthetic benchmarks and real-world data, fragility to distribution shifts, and failure in compositional generalization, collectively questioning its...

    arxiv.org/abs/2605.30015 · PDF

  42. 42

    MIC: Maximizing Informational Capacity in Adaptive Representations via Isotropic Subspace Alignment

    Dang Hong Nguyen, Nhi Ngoc-Yen Nguyen, Huy-Hieu Pham

    cs.LG · cs.CL

    Although multi-scales representation learning enables elastic-dimension embeddings, nested subspaces often suffer from dimensional redundancy and spectral collapse. To address this, we introduce MIC, a framework that optimizes the geometric landscape of multi-granular embeddings through isotropic subspace alignment. MIC employs Soft Collapse Regularization (SCR) to mitigate redundancy between prefix and residual subspaces via...

    arxiv.org/abs/2605.29987 · PDF

  43. 43

    Improving Adversarial Robustness of Attribution via Implicit Regularization

    Amir Mehrpanah, Matteo Gamba, Hossein Azizpour

    cs.LG · cs.CV

    The adversarial robustness of attributions is a fundamental requirement for reliable explainability in deep learning, yet existing approaches typically rely on computationally expensive explicit regularization. In this work, we show that attribution robustness can arise implicitly from the learning dynamics of standard stochastic gradient descent. We theoretically motivate this effect through connections between parameter-space and...

    arxiv.org/abs/2605.29983 · PDF

  44. 44

    A Fully Convolutional Approach to Denoising Structural Dynamics Data from X-Ray Photon Correlation Spectroscopy

    Nisar Nellikunnummel, Andi Barbour, Lutz Wiegart, Tatiana Konstantinova, Anthony DeGennaro

    cs.LG · eess.SP

    We present a fully convolutional denoising autoencoder (FC-DAE) for denoising two-time intensity-intensity correlation functions ($C_2$) in X-ray photon correlation spectroscopy (XPCS). Unlike conventional denoising autoencoders that are typically restricted to fixed input sizes, the FC-DAE accepts inputs of arbitrary dimensions while preserving correlation structures across diverse dynamical regimes. The model is trained using experimentally...

    arxiv.org/abs/2605.29975 · PDF

  45. 45

    From Short Histories to Long Futures: Horizon-Aware Graph Neural Networks for Long Horizon Forecasting

    Zesheng Liu, Maryam Rahnemoonfar

    cs.LG

    Accurate long-range prediction of geophysical systems is difficult due to strongly nonlinear dynamics, the high computational cost of full-physics simulations, and the error accumulation that arise when one-step autoregressive surrogates are rolled out over decades. Deep neural network can serve as efficient emulators, but most are trained only for next-step prediction and often drift or become unstable as the forecast horizon grows. We...

    arxiv.org/abs/2605.29952 · PDF

  46. 46

    CLUBench: A Clustering Benchmark

    Feng Xiao, Dazhi Fu, Chris Ding, Jicong Fan

    cs.LG

    Clustering is a fundamental problem in data science with a long-standing research history, yielding numerous insightful algorithms. Despite this progress, a systematic and large-scale empirical evaluation that jointly considers conventional algorithms, deep learning-based methods, and recent foundation model-based clustering remains largely absent, leading to limited guidance on algorithm selection and deployment. To address this gap, we...

    arxiv.org/abs/2605.29933 · PDF

  47. 47

    Treatment-Conditioned Diffusion for Forecasting Neurodegenerative Disease Progression

    Danylo Boiko, Viktoriia Mishkurova

    cs.LG · cs.CV

    Forecasting the progression of neurodegenerative diseases, such as Parkinson's disease, is essential for effective long-term planning and personalized therapeutic intervention. Existing systems typically produce scalar clinical scores that ignore the rich structure of longitudinal neuroimaging, while traditional generative approaches suffer from a loss of anatomical details and blurring subtle progression patterns. To address this, we...

    arxiv.org/abs/2605.29932 · PDF

  48. 48

    A Triple-Modal Contrastive Learning Framework with Sequence, Graph, and 3D Features for Drug-Target Interaction Prediction

    Le Xu, Xi Zhang, Dan Luo, Ting Wang, Xuan Lin

    cs.LG

    Accurate prediction of drug-target interactions (DTI) is critical for drug discovery. Existing methods often rely on single-modal representations (e.g., sequences or graphs) or combine only two modalities, overlooking 3D structural features. To address this challenge, we propose TriMod-DTI, a triple-modal contrastive learning framework that incorporates 1D sequences, 2D graphs, and 3D structures of drugs and proteins, obtaining the universal...

    arxiv.org/abs/2605.29926 · PDF

  49. 49

    Midpoint Generative Models

    Daniil Shlenskii, Nikita Gushchin, Lev Novitskiy, Dmitry V. Dylov, Alexander Korotin

    cs.LG

    We introduce Midpoint Generative Models (MGM), a principled framework for training one-step generative models. MGM is based on a simple symmetry of Flow Matching with linear interpolation: when the two endpoint distributions coincide, the corresponding drift field vanishes at the midpoint time, $t=1/2$. We show that the norm of this field defines a valid discrepancy between distributions, which we call the Midpoint Divergence. We extend this...

    arxiv.org/abs/2605.29920 · PDF

  50. 50

    Reducing Experimental Testing in Space Propulsion Film Cooling Analyses by Pixelwise Generative Image Interpolation

    Adam T. Müller, Philipp J. Teuffel, Konstantin Manassis, Nicolaj C. Stache

    cs.LG · cs.CV

    We propose a machine learning approach for image regression from sparse experimental measurements. We show the application of the proposed method on film cooling studies in propulsion system development, aiming to reduce the need for extensive physical testing. Our method employs a lightweight feed-forward neural network with positional encoding to generate images conditioned by input parameters. Validated on real and synthetic data, it...

    arxiv.org/abs/2605.29911 · PDF

  51. 51

    Plan, Don't Pose: Long Composite Motion Generation with Text-Aligned BFM

    Nikolay Shvetsov, Maksim Bobrin, Nazar Buzun, Dmitry V. Dylov

    cs.LG

    Text-to-motion (T2M) generation has broad applications in character animation, virtual avatars, and human-robot interaction. Existing methods typically generate pose trajectories or motion tokens directly from language, forcing a single model to handle semantic interpretation, long-horizon structure, and low-level physical realization. This coupling makes them costly and often unreliable for long, compositional, or semantically dense prompts....

    arxiv.org/abs/2605.29906 · PDF

  52. 52

    OVA-IB: One vs All Information Bottleneck for Multi-Modal Alignment

    Tianchao Li, Shujian Yu, Xinrui Zu, Zhaolong Wei, Jeremy Gummeson, Jack C. P. Cheng, Robert Jenssen

    cs.LG · cs.IT

    Contrastive learning is effective for aligning paired views or modalities, but alignment beyond two modalities remains non-trivial and comparatively underexplored. Pairwise CLIP-style losses decompose multi-modal alignment into independent two-way comparisons and therefore do not explicitly model higher-order dependencies among multiple modalities. Recent beyond-pairwise objectives approach this problem from statistical or geometric...

    arxiv.org/abs/2605.29900 · PDF

  53. 53

    LaRA: Layer-wise Representation Analysis for Detecting Data Contamination in RL Post-Training

    Minju Gwak, Minseo Kwak, Dongseok Lee, Guijin Son, Alan Ritter, Jaehyung Kim

    cs.LG · cs.AI

    Reinforcement learning (RL) post-training has shown to improve reasoning in large language models (LLMs). However, there has been little exploration on the problem of data contamination in RL post-training, potentially undermining generalization and evaluation reliability of the training process itself. Existing detection methods primarily rely on output-level signals such as likelihood or entropy, which become unreliable for RL-trained...

    arxiv.org/abs/2605.29888 · PDF

  54. 54

    Open Problem: Separating Geometric and Algorithmic Compression via Cayley-Table Completion

    Dongsung Huh

    cs.LG · cond-mat.dis-nn · math.OC · math.RT · stat.ML

    Modern statistical learning theory and deep learning characterize generalization primarily in terms of continuous capacity control (e.g., norm-based regularization, margin maximization, low-rank bias). While highly successful in continuous domains, deep learning consistently fails to extrapolate exact algorithmic or discrete algebraic rules, reflecting a missing inductive bias toward algorithmic complexity minimization. We propose the...

    arxiv.org/abs/2605.29885 · PDF

  55. 55

    STAP: A Shuffle-Tokenized App Predictor with Ultra Long Context for Vocabulary-Free Mobile App Prediction

    Chengyu Fan, Hang Liu

    cs.LG

    Predicting the next mobile application a user will launch is essential for intelligent device resource management and proactive assistance. Existing models rely on fixed app vocabularies, which prevents them from generalizing across different app ecosystems. Many also depend on user-specific knowledge, which complicates deployment in cold start scenarios. We propose STAP, a Transformer-based model that eliminates the need for a fixed...

    arxiv.org/abs/2605.29863 · PDF

  56. 56

    ESPO: Early-Stopping Proximal Policy Optimization

    Zihang Li, Rui Zhou, Yingcheng Shi, Wenhan Yu, Zhewen Tan, Zixiang Liu, Zeming Li, Binhua Li, Yongbin Li, Tong Yang,...

    cs.LG · cs.AI

    When a large language model under reinforcement learning commits a wrong reasoning step early in a trajectory, standard algorithms force it to keep generating until the maximum horizon, spending compute on tokens that never receive positive reward and polluting advantage estimates with post-failure noise. We propose ESPO (Early-Stopping Proximal Policy Optimization), which detects trajectory failure on-the-fly and terminates rollouts early....

    arxiv.org/abs/2605.29860 · PDF

  57. 57

    Feedback-to-Rubrics: Can We Learn Expert Criteria from Inline Comments?

    Kotaro Yoshida, So Kuroki, Yuki Imajuku, Taishi Nakamura, Ryunosuke Iwai, Haruki Goda, Takuya Akiba

    cs.LG

    Large language models (LLMs) are increasingly used for writing and review support, but their usefulness depends on context-dependent criteria, such as expert preferences or organization-specific conventions, that are often tacit, undocumented, and difficult to elicit directly. We propose a problem setting for learning reusable natural-language rubrics from accumulated inline comments on artifacts such as human-written or LLM-generated drafts....

    arxiv.org/abs/2605.29857 · PDF

  58. 58

    MIRAGE: Adaptive Multimodal Gating for Whole-Brain fMRI Encoding

    Abdulkadir Gokce, Badr AlKhamissi, Martin Schrimpf

    cs.LG

    Recent progress in task-optimized neural networks has established encoding models as a powerful tool for predicting brain responses to naturalistic stimuli, yet most existing approaches rely on unimodal representations. The emergence of omni-modal foundation models and rich multimodal neural datasets enables encoding models that jointly integrate visual, auditory, and linguistic information across subjects. We introduce MIRAGE, a brain...

    arxiv.org/abs/2605.29850 · PDF

  59. 59

    HARP: Hadamard-Preconditioned Adaptive Rotation Processor for Extreme LLM Quantization

    Artur Zagitov, Gleb Molodtsov, Aleksandr Beznosikov

    cs.LG · cs.AI

    Post-training quantization (PTQ) is essential for deploying LLMs under memory and bandwidth constraints. However, extreme low-bit quantization remains highly sensitive to activation outliers and anisotropic weight curvature. Existing incoherence-based PTQ methods mitigate this issue with fixed randomized Hadamard transforms (RHTs), which improve quantization robustness but cannot adapt the rotated basis to the layer, calibration distribution,...

    arxiv.org/abs/2605.29843 · PDF

  60. 60

    CB-SLICE: Concept-Based Interpretable Error Slice Discovery

    Yael Konforti, Mateo Espinosa Zarlenga, Elaf Almahmoud, Mateja Jamnik

    cs.LG · cs.AI · stat.ML

    Despite strong average-case performance, deep learning models often exhibit systematic errors on specific population groups, known as error slices. Identifying these groups and the root causes of their failures is critical for model debugging and bias mitigation. However, existing error Slice Discovery Methods (SDMs) typically generate explanations disconnected from the model's inference process, thus only approximating the underlying error...

    arxiv.org/abs/2605.29836 · PDF

  61. 61

    Open World Autoencoding Drift Detection with Novel Class Recognition in Tabular Non-stationary Data Streams

    Joanna Komorniczak

    cs.LG

    Data stream processing has become a landmark in modern machine learning applications, with concept drifts and novel class appearances posing the primary challenges faced by sophisticated recognition methods. This work proposes an unsupervised concept drift detection method that identifies shifts in known class distributions based on the reconstruction errors of an autoencoder, while also enabling the recognition of novel class samples through...

    arxiv.org/abs/2605.29834 · PDF

  62. 62

    When Do Graph Foundation Models Transfer? A Data-Centric Theory

    Jiajun Zhu, Ying Chen, Peihao Wang, Yixuan He, Pan Li, Aditya Akella, Zhangyang Wang

    cs.LG

    Graph foundation models (GFMs) aim to reuse a single backbone across diverse graph domains, yet their transfer is often uneven and can exhibit negative transfer. While most prior work improves transfer through architectural or adaptation choices, we ask a data-centric question: which properties of two graph domains determine how much a fixed representation model changes its outputs? Using a graphon-based continuous limit for dense graphs, we...

    arxiv.org/abs/2605.29828 · PDF

  63. 63

    The Interplay Between Interpolation and Aggregation in Regression: Optimal Sample Complexity

    Mikael Møller Høgsgaard, Kasper Green Larsen, Liang-Yu Zou

    cs.LG

    This work investigates theoretically the interplay between interpolation and aggregation in regression. We establish that the $γ$-graph dimension characterizes learnability for a broad class of natural aggregation procedures. Furthermore, we prove that an extremely simple aggregation procedure, combining three interpolating hypotheses via the median, is optimal among all these aggregation procedures, and is strictly more powerful than proper...

    arxiv.org/abs/2605.29819 · PDF

  64. 64

    Gated Graph Attention Networks with Learnable Temperature

    Zhongtian Ma, Hao Wu, Yexin Zhang, Qiaosheng Zhang, Zhen Wang

    cs.LG

    Graph attention networks learn neighbor importance through data-dependent coefficients, but standard layers lack explicit control over unreliable feature dimensions and use fixed sharpness of attention coefficient distributions. This paper proposes gated graph attention and learnable temperature for common graph attention mechanisms. Gated graph attention filters feature or message responses to reduce the influence of unreliable dimensions,...

    arxiv.org/abs/2605.29803 · PDF

This edition is part of The Daily Abstract — cs.LG archive. Subscribe to receive these in your inbox each morning, automatically translated to Spanish, with reply-to-PDF: arxivdaily.ignorelist.com.

Colophon Set in Georgia, with system sans for interface chrome and a monospaced stack for code and paper identifiers. Sole accent: amber #D99C5E. Built and served on an always-free VM. The masthead is set 14% letterspaced because newspapers do that and it works.