cs.LG · 2026-09-03 · No. 104

Machine Learning, 2026-09-03.

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

01 — The papers

59 entries
  1. 01

    A Common Measure of Communication for Speech Brain-Computer Interfaces

    Dulhan Jayalath, Benjamin Ballyk, Oiwi Parker Jones

    cs.LG · q-bio.NC

    Speech brain-computer interfaces (speech BCIs) translate neural activity into language, offering a path towards restoring speech for people with paralysis and, more broadly, enabling new forms of natural human-computer interaction. Despite this promise, the field lacks a common measure of progress because systems use different datasets, recording methods, types of speech, and vocabularies, so their reported scores are rarely comparable....

    arxiv.org/abs/2609.02887 · PDF

  2. 02

    Graph Machine: Towards Better Pretraining via Edges

    Lintai Hou

    cs.LG

    We introduce the Graph Machine (GM), an architecture that maintains an $O(n)$-sized state and accesses it through sparse, dynamic routing. Unlike methods with fixed-size states or sparse but static routing, GM preserves $O(n)$ complexity in its sparse layers without restricting the potentially accessible state size to $O(1)$. Instead, GM uses edges - pointer-like objects updated differentiably by a referral mechanism resembling pointer...

    arxiv.org/abs/2609.02881 · PDF

  3. 03

    The Implications of Linguistic Illegibility for LLM Security

    James Mickens

    cs.LG · cs.CR

    LLMs are trained to generate natural language. However, various strands of evidence indicate that an LLM's externalized linguistic outputs and mechanistically-extracted linguistic features can be an unreliable lens for understanding internal model computation. We introduce the term ``linguistic illegibility'' to broadly refer to scenarios in which an LLM's externalized or mechanistically-probed language artifacts fail to represent how the...

    arxiv.org/abs/2609.02852 · PDF

  4. 04

    Post-Training Language Models for Gold-Medal Performance in Coding Competitions

    Aleksander Ficek, Sean Narenthiran, Mehrzad Samadi, Somshubra Majumdar, Boris Ginsburg

    cs.LG · cs.AI · cs.CL · cs.MA · cs.SE

    Competitive programming has become a key test of large language model reasoning, with international competitions such as IOI and ICPC representing its most challenging settings. We present an end-to-end specialization pipeline combining large-scale problem curation, synthetic reasoning traces, supervised fine-tuning (SFT), and reinforcement learning (RL). Using 22,000 curated problems, we train Nemotron-3-Nano-CC (30B-A3B) with SFT and RL and...

    arxiv.org/abs/2609.02849 · PDF

  5. 05

    UE5M3 FP4 Block Scaling for Stable Language Model Pretraining

    Robert Hu, Carlo Luschi, Paul Balanca

    cs.LG

    Stable 4-bit floating-point (FP4) pretraining is difficult because the E2M1 payload represents only a narrow range of magnitudes. NVIDIA's Transformer Engine \nv{} recipe addresses this with current-tensor scaling, a randomized Hadamard transform (RHT), and bfloat16 (BF16) final layers, adding work outside the FP4 matrix multiplications. We instead pair E2M1 payloads with unsigned E5M3 (\ue{}) block scales. Their wider range permits periodic...

    arxiv.org/abs/2609.02846 · PDF

  6. 06

    Cliff: Learning Process Rewards from the First Mistake

    Peixuan Han, Runhui Wang, Ketan Ramaneti, Jie Hao, Gerald Friedland, Chris Kong

    cs.LG

    Reinforcement learning with verifiable rewards (RLVR) has emerged as a powerful paradigm for large language model (LLM) post-training, but its reliance on coarse outcome rewards leads to limited guidance on intermediate reasoning processes. Existing approaches such as process reward modeling and on-policy distillation introduce additional constraints, such as reliance on a specialized reward model or assuming identical reasoning patterns...

    arxiv.org/abs/2609.02817 · PDF

  7. 07

    Do Tabular Foundation Models Know Physics? Contamination, Units, and the Deterministic Limit

    Wassim Tenachi, Yashar Hezaveh, Laurence Perreault Levasseur, Pierre-Luc Bacon

    cs.LG · astro-ph.IM

    Tabular foundation models (TFMs) learn to fill in tables the way language models fill in text, and tables are arguably the format in which most physical measurement arrives. Did they learn any physics in the process? They are Bayesian by construction, so the question is what their prior contains. We probe it directly, evaluating four of them (TabPFN-3, TabICLv2, TabDPT and Real-TabPFN-2.5) against six baselines on datasets sampled from 316...

    arxiv.org/abs/2609.02766 · PDF

  8. 08

    LoRA-TSD: Tangent-Space Spectral Descent for LoRA via Muon-Style Updates

    Dmitrii Andriianov, Andrey Veprikov, Aleksandr Beznosikov

    cs.LG

    Low-rank adaptation (LoRA) is the standard way to fine-tune large models, yet when its two factors are trained independently, the update ignores the geometry of the low-rank weight change it induces. We introduce LoRA-TSD, an optimizer that treats every LoRA step as a tangent vector of the fixed-rank matrix manifold and takes the spectral-norm steepest-descent step of Muon inside that tangent space, mapping the result back to the factors...

    arxiv.org/abs/2609.02734 · PDF

  9. 09

    H3DNAS: Hardware-Aware ONNX-Native 3D Point Cloud Model Compression

    Anchit Mulye, Rhythm Baghel, Sujay Kumar Ingle, Hardik Jain

    cs.LG · cs.AR · cs.NE

    Deploying 3D point cloud models on edge hardware such as the NVIDIA Jetson Orin Nano is severely constrained by compute and memory budgets. Existing compression methods require access to the model's original source code, rendering them inapplicable to the Open Neural Network Exchange (ONNX) binaries commonly distributed by vendors and model repositories. We present \textbf{H3DNAS}, a hardware-aware model compression framework that operates...

    arxiv.org/abs/2609.02684 · PDF

  10. 10

    Unfolding the Leech Lattice: Fused Multi-Shell Decoding and VRAM Layouts for 2-Bit LLM Weights

    Pier-Jean Malandrino

    cs.LG

    Leech-lattice vector quantization holds the strongest reported 2-bit quality under its own evaluation protocol. Its kernel decodes one shell; we found no implementation of the multi-shell decoder the rate requires. This paper supplies one and measures its serving cost for decode-phase GEMV at batch 1. First, a serving path for the full 301-class codebook: an offline expansion into GPU layouts and a fused dequantize-plus-matvec kernel reading...

    arxiv.org/abs/2609.02652 · PDF

  11. 11

    Differentiable Electricity-Market Clearing for Gradient-Based Planning

    Luca Mungo, Maarten P. Scholl, Arnau Quera-Bofarull

    cs.LG · cs.CY

    Planning a large data center is difficult because a facility big enough to matter changes the electricity prices it will pay. Those prices are set by market clearing, a constrained optimization problem solved anew in every operating condition. However, simulating the market tells a planner how a candidate plan performs but not how to improve it. Here we treat market clearing as a differentiable optimization layer: each forward pass solves the...

    arxiv.org/abs/2609.02646 · PDF

  12. 12

    Oracle, will I ever learn? A study of prediction convergence and complementarity across link prediction models

    Guillaume Méroué, Fabien Gandon, Pierre Monnin

    cs.LG

    Knowledge graphs have become an important source of structured knowledge for Web applications, including search, question answering, and recommender systems. In these applications, link prediction can serve either as a prediction task itself or as a means to enrich incomplete knowledge graphs for downstream tasks. Interestingly, different link prediction models, or even different training runs of the same model, can produce substantially...

    arxiv.org/abs/2609.02638 · PDF

  13. 13

    Source Distribution Estimation by Posterior Averaging

    Trung-Dung Hoang, Lisa M. Koch

    cs.LG

    Simulation-based science often requires a distribution over simulator parameters whose push-forward reproduces a set of real observations: this is the source distribution estimation (SDE) problem. Existing methods fit the source against a likelihood surrogate trained once from a fixed proposal prior. Their objective is therefore stated only in terms of the surrogate instead of the true simulator, which may fail for inaccurate areas in...

    arxiv.org/abs/2609.02622 · PDF

  14. 14

    Online Reinforcement Learning in the Met Office Unified Model through Distributed Model-Agent Coupling

    Pritthijit Nath, Sebastian Schemm, Peter Haynes, Emily Shuckburgh, Mark Webb

    cs.LG

    Machine-learnt corrections can complement numerical weather prediction only if they adapt to the evolving model state while preserving dynamical consistency and numerical stability. To test this within a global forecasting model, we couple the Met Office (UKMO) Unified Model (UM) with distributed RL agents through rank-local tensors. A DDPG actor shares weights across the 70 vertical model levels of each atmospheric column and applies bounded...

    arxiv.org/abs/2609.02566 · PDF

  15. 15

    ProbeMatchDTI: Probe-Driven Multi-Scale Biochemical Pattern Matching for Drug-Target Interaction Prediction

    Quan Hao, Mengyue Fan, Zifan Dong, Youru Li, Jianduo Zhao, Lechuan Xu, Hao Zhang, Fei Xia, Jigang Wang, Chong Qiu,...

    cs.LG · cs.AI

    Drug-target interaction (DTI) prediction is an important task in AI-driven drug discovery. Although recent biochemical representation learning methods have improved DTI prediction, their passive feature aggregation tends to favor dominant molecular patterns while suppressing weak yet binding-relevant signals, such as functional groups and residue-context patterns, limiting the modeling of multi-scale biochemical correspondences. To address...

    arxiv.org/abs/2609.02549 · PDF

  16. 16

    Learn from Whoever Is Right: Answer-Verified Multi-Teacher Distillation for Multi-Domain LLMs

    Xixiang He, Xingming Li, Baiqi Wu, Qiyao Sun, Xuanyu Ji, Ao Cheng, Qingyong Hu

    cs.LG · cs.AI

    Modern large language models (LLMs) rely on reinforcement learning to build strong capabilities in individual domains, but integrating those capabilities into a single deployable model remains challenging. By routing each sample to the teacher whose domain matches it, existing approaches let a domain label decide which teacher provides supervision. However, domain expertise holds only on average: the matched teacher is not always correct on a...

    arxiv.org/abs/2609.02548 · PDF

  17. 17

    TrajMind: Chaining Role-Specialized LoRAs for Fast-and-Slow Collective Trajectory Anomaly Diagnosis

    Jiahao Wu, Zhenqun Yang, Chen Jason Zhang, Qing Li

    cs.LG

    Diagnosing collective anomalies from urban trajectories is increasingly important for traffic governance, as it reveals what happened, who was involved, and where and when the event occurred. Existing detectors efficiently produce scores or labels, whereas vision--language pipelines provide richer semantics; neither couples verifiable diagnosis with low-latency monitoring. The central challenge is to recognize collective patterns and recover...

    arxiv.org/abs/2609.02540 · PDF

  18. 18

    A Comparative Study of Graph Representations for GNN-Based Power Grid Control in L2RPN

    Adrian Degenkolb, Qiong Huang, Benjamin Schäfer

    cs.LG

    Graph construction is a critical but underexamined design choice in deep reinforcement learning for power grid control. We present a controlled experimental comparison of different graph representations, including physical topology, electrical-sensitivity, and hybrid variants for topology control in the Learning to Run a Power Network (L2RPN) environment. Our findings indicate that matching graph complexity to task granularity is more...

    arxiv.org/abs/2609.02538 · PDF

  19. 19

    Spectral Initialization and Scheduled Graph Smoothness for Uncertain Knowledge Graph Completion

    Md Abrar Jahin, Taufikur Rahman Fuad, Jay Pujara, Craig A. Knoblock

    cs.LG · cs.AI

    Uncertain knowledge graphs (UKGs) extend knowledge graphs by assigning each triple a continuous confidence score. Since most possible triples lack observed confidences, recent methods rely on semi-supervised learning to generate pseudo-labels. These methods initialize entity embeddings without using the confidence-weighted graph, discarding its global community and hub structure. We introduce QUEST, which adds no trainable parameters to the...

    arxiv.org/abs/2609.02519 · PDF

  20. 20

    Rethinking the Teacher-Student Framework for Test-Time Adaptation

    Damian Sójka, Marc Masana, Bartłomiej Twardowski, Sebastian Cygert

    cs.LG

    Test-Time Adaptation (TTA) has recently emerged as a promising strategy that allows the adaptation of pre-trained models to changing data distributions at deployment time, without access to any labels. To mitigate error accumulation, researchers have widely adopted the teacher-student framework, though its long-term stability is often taken for granted. In this work, we challenge the common strategy of setting the teacher weights to an...

    arxiv.org/abs/2609.02507 · PDF

  21. 21

    RINSE: Robust Target-Time Normality Estimation for Zero-Shot Graph Anomaly Detection

    Taufikur Rahman Fuad, Md Abrar Jahin, Amir Hussain

    cs.LG · cs.AI

    Zero-shot graph anomaly detection seeks to deploy a detector trained on source graphs to unseen, unlabeled targets, yet domain shift can make source-derived notions of normality unreliable. We introduce RINSE (Robust Iterative Normality Self-Estimation), a gradient-free target-time framework that keeps the source-trained detector fixed while sequentially estimating target normality, representation calibration, and evidence reliability from...

    arxiv.org/abs/2609.02497 · PDF

  22. 22

    DeepAffinity: Long-Term Aspect Preference Prediction in eCommerce using Small Language Models

    Yotam Eshel, Guy Hadad, Guy Feigenblat, Yuri M. Brovman, Matt Gearhart, Bracha Shapira

    cs.LG · cs.AI

    We explore predicting eCommerce user preferences for product aspects such as brand, size, and color - a task we define as Aspect Affinity. Solving this task improves customer understanding and enables fine-grained personalization in recommendation, search, and marketing. We frame Aspect Affinity as a temporal prediction task: forecasting a users future aspect choices from their time-ordered interaction history, capturing long-term preferences...

    arxiv.org/abs/2609.02468 · PDF

  23. 23

    Scalable Kronecker-Fisher Approximation: Efficient Hessian Analysis for Billion-Parameter Language Models Compression

    Viacheslav Yusupov, Daria Cherniuk, Evgeny Frolov

    cs.LG · cs.AI · cs.CL

    In this paper, we propose a scalable Kronecker-based approximation that captures cross-layer interactions without storing the entire Fisher matrix, enabling practical Hessian analysis for billion-parameter networks where full computation is infeasible. Our approach reveals consistent vulnerability patterns: value projection layers exhibit the highest sensitivity and strongest cross-layer correlations across multiple model families, while...

    arxiv.org/abs/2609.02451 · PDF

  24. 24

    CACTUS: Mask-Guided Semantic Clean-Label Backdoors in Decentralized Federated Learning

    Chao Feng, Burkhard Stiller

    cs.LG · cs.DC

    Semantic triggers in federated learning (FL) can be less conspicuous than synthetic patches, but sample-dependent placement may weaken backdoor implantation across aggregation rounds. This challenge is compounded in decentralized FL (DFL), where topology-dependent peer aggregation repeatedly mixes local models. CACTUS converts label-consistent semantic pairs into target-directed representation shifts. Mask-guided, modality-specific operators...

    arxiv.org/abs/2609.02450 · PDF

  25. 25

    Towards One-for-All Robustness Across a Continuum of Threat Levels

    Zhichao Hou, Xiaorui Liu

    cs.LG · cs.AI

    Adversarially robust models often overfit to a specific attack budget, necessitating multiple specialized models for diverse and dynamic adversarial environments, a strategy that becomes fundamentally intractable as the threat space grows. This raises an open challenge: can we achieve strong robustness across a continuum of threat levels within a single model? We propose the Threat Conditional Network (TCN), grounded in a representation...

    arxiv.org/abs/2609.02440 · PDF

  26. 26

    IFW-BLS: Dual-Robust Broad Learning System with Intuitionistic Fuzzy Wave Loss

    Mushir Akhtar, M. Tanveer

    cs.LG

    Broad Learning System is an efficient randomized learning model that expands network width through feature and enhancement nodes and estimates the output weights without deep backpropagation. Its standard least-squares training, however, is vulnerable in two different ways: (i) large residuals caused by noise, outliers, or corrupted labels can dominate the objective, and (ii) all samples are treated as equally reliable even when some lie in...

    arxiv.org/abs/2609.02422 · PDF

  27. 27

    Coverage, Not Targeting: A Structural Regime in Multi-Turn Agent Credit Assignment

    Chenyu Zhou, Qiliang Jiang, Shuning Wu, Xu Zhou

    cs.LG · cs.AI

    Multi-turn agentic RL increasingly treats credit assignment as a targeting problem: given a terminal verifiable reward, per-turn methods localize credit onto the turns that mattered. We identify the structural quantity that predicts when this is the right move, the verifier information density V_d = k/C (the fraction of an agent's C-step causal chain whose per-turn correctness the verifier exposes), and show that terminal-state verifiers sit...

    arxiv.org/abs/2609.02417 · PDF

  28. 28

    Evidence for Shared Routing Geometry and Dynamics in Sparse Mixture-of-Experts

    Kirill Labzin, Stepan Kulibaba, Artem Dzhalilov, Artem Gorokhov

    cs.LG · cs.AI

    Sparse mixture-of-experts (MoE) models use an independently parameterized router at each sparse layer to select experts for every token. Prior work has shown that routing decisions across depth can often be predicted from earlier routing signals, suggesting that routing is not fully independent across layers. However, the structure behind this predictability remains unclear. In this work, we provide evidence that routing-relevant states...

    arxiv.org/abs/2609.02404 · PDF

  29. 29

    Percolation Dynamics in Optimization : Variance Cascades and Discrete Scale Invariance

    Sai Niranjan Ramachandran, Suvrit Sra

    cs.LG · cond-mat.dis-nn · cond-mat.stat-mech · cs.AI

    We study the dynamics of Stochastic Gradient Descent (SGD), which is known to steer deep neural networks toward invariant sets that correspond to simpler subnetworks. How this steering unfolds over time remains poorly understood. We answer this by modeling the stochastic gradient flow (SGF) as a percolation process, in which architectural symmetries force subnetworks to merge in discrete simultaneous blocks rather than one at a time. These...

    arxiv.org/abs/2609.02373 · PDF

  30. 30

    AGI Maze Prediction Datasets: A Compact Benchmark for Learning World Dynamics with Transformers

    Alexey Potapov

    cs.LG · cs.AI

    World modeling requires a predictive model to maintain and update an internal state adequate for reasoning about the consequences of actions. We introduce the AGI Maze Prediction Datasets and Benchmark, a lightweight controlled testbed for studying this capability in Transformers and other predictive models. Derived from procedurally generated, stateful grid worlds, the benchmark comprises per-step transition prediction, fixed-horizon state...

    arxiv.org/abs/2609.02339 · PDF

  31. 31

    What Is Worth Representing? Representational Empowerment for Continual Model Construction

    Fei Dai, Hanqi Zhou, Alison Gopnik, Charley Wu

    cs.LG · cs.AI

    The first problem of modeling the world is not just estimating the right parameters or causal structure, but deciding what should be represented at all. We frame this problem as continual model construction: an agent maintains an environment-specific model M of an inaccessible world W and curates a persistent library L of reusable representational elements across environments. We propose Representational Empowerment (RepEmp) to score...

    arxiv.org/abs/2609.02322 · PDF

  32. 32

    Bayes-Optimal BER and AUC: Estimation and Evaluation of Estimators

    Ryota Ushio, Takashi Ishida, Masashi Sugiyama

    cs.LG · stat.ML

    A fundamental quantity in machine learning is the optimal performance achievable by any model on a given task. Estimating this quantity allows us to distinguish the irreducible part of the error from a deficiency of the model, telling us how much room for improvement remains. Recent work has shown that the Bayes error, or equivalently the optimal accuracy, can be estimated from soft labels in binary classification. However, accuracy is often...

    arxiv.org/abs/2609.02304 · PDF

  33. 33

    SEAL: Reinforcing Global Safety in Mixture-of-Experts through Shared Expert ALignment

    Qingyu Meng, Yiwei Zha, Jiahuan Pei, Koen Hindriks, Herbert Bos, Min Chen

    cs.LG · cs.AI · cs.CR

    Mixture-of-Experts (MoE) is a scaling architecture for large language models that activates only a small subset of expert modules per token, enabling massive parameter growth with nearly constant computation. Recent Hybrid MoE architecture adds \textit{shared experts} to capture consistently useful representations, further improving stability and generalization. MoE now powers many flagship open-source and commercial models, yet remains...

    arxiv.org/abs/2609.02293 · PDF

  34. 34

    Entangled Representations Amplify Collateral Damage in Unlearning

    Evžen Wybitul, Tim G. J. Rudner, Christian Schroeder de Witt

    cs.LG · cs.CL

    A long-held intuition in interpretability research is that representational entanglement, the sharing of structure between knowledge domains in a neural network, makes unlearning harder. While the intuition is widespread, it has never been directly tested in a controlled experiment. We present a way to do so: by repurposing Selective Gradient Masking (SGTM), we train a suite of six 254M-parameter language models on English Wikipedia with...

    arxiv.org/abs/2609.02285 · PDF

  35. 35

    CAPTURE: Disentangling Preference Drift from Memory Poisoning in Personalized LLM Agents

    S M Asif Hossain, Ruksat Khan Shayoni, Md Kishor Morol

    cs.LG

    Personalized language agents use persistent memory to adapt to users over time, but the same mechanism creates an attack surface. When new information conflicts with stored preferences, an agent must distinguish genuine preference drift from temporary context shifts, ambiguity, or adversarial memory poisoning. We formulate this problem as a continuous-time partially observable decision process over a latent user state and show why rules based...

    arxiv.org/abs/2609.02265 · PDF

  36. 36

    Similarity-Aware Personalized Federated Learning in Heterogeneous Environments

    Arun Kumar A, Sunil Gupta, Dang Ngyuen, Bao Duong, Dat Phan Trong

    cs.LG

    Federated Learning (FL) allows decentralized clients to train models collaboratively while preserving data privacy. However, distribution mismatch across clients often leads to poor global generalization and degraded local client-level performance. In such scenarios, some of the clients with their local models trained solely on local data may perform better than the globally learnt model, thus nullifying the benefits of collaborative...

    arxiv.org/abs/2609.02241 · PDF

  37. 37

    Recursive Value Learning for Long-Horizon Offline Goal-Conditioned RL

    Hyeonseong Jeon, Youngwoon Lee

    cs.LG · cs.RO

    Scaling offline goal-conditioned reinforcement learning (GCRL) to long-horizon tasks is difficult because (1) long-range value learning depends on shorter-range estimates that may still be inaccurate, and (2) max-based value backups can amplify overestimation through repeated propagation. We propose DCRL (Divide-and-Conquer RL), which recursively decomposes each trajectory segment into a balanced binary tree and trains the values from leaves...

    arxiv.org/abs/2609.02237 · PDF

  38. 38

    SMart: A Multi-source Multi-phase Time Series Representation Transfer Framework

    Fang He, Wang-chien Lee

    cs.LG · cs.AI

    Time series representation learning (TSRL) has attracted growing research interests in recent years. Two recent explorations in TSRL are: i) exploiting a transformer-based framework to learn time series; ii) instead of using only the targeted dataset, borrowing time series from other datasets to to facilitate representation transfer. While these two explorations are shown effective, the self-supervised time series recovery task in (i) and the...

    arxiv.org/abs/2609.02203 · PDF

  39. 39

    Learning the Constitutive Behavior of Materials via Neural Operators and Causal Attention: Case Studies in Plasticity and Damage

    Rishabh Arora, Lisa Scheunemann, Tim Brepols, Shahed Rezaei

    cs.LG · cs.CE

    Classical constitutive modeling of path-dependent inelastic materials relies on internal state variables whose evolution equations must be postulated based on domain knowledge and calibrated against experimental data. However, in many practical settings, the relevant internal variables are typically not measurable in experiments, and the constitutive response must be inferred entirely from measured strain-stress data without any prior...

    arxiv.org/abs/2609.02194 · PDF

  40. 40

    DMRL: Document-Mediated Reinforcement Learning for Skill Optimization in Advertising Recommendation

    Wei Zhang, Hongji Li, Song Sun, Peng Yu, Xue Yang, Lei Zhao, Peng Jiang

    cs.LG

    Advertising recommendation requires continuously tuning complex system parameters while balancing commercial returns and user experience. Recent work has introduced large language models (LLMs) with skill documents to assist this labor-intensive process, but skill optimization remains largely prompt-driven, lacking a principled mechanism to attribute rewards to specific document edits. To address this limitation, we propose Document-Mediated...

    arxiv.org/abs/2609.02170 · PDF

  41. 41

    GeoSPRINT: Geometric Redundancy-Aware Step Pruning for Inference in Diffusion Trajectories

    Arpita Joshi

    cs.LG · cs.AI

    Diffusion models achieve high sample quality but remain expensive at inference time because sampling requires many sequential neural function evaluations (NFEs). Existing acceleration methods either use fixed step-skipping schedules, adapt step sizes based on local numerical error, or require additional training. We introduce GeoSPRINT (Geometric Step Pruning for Inference in Trajectories), a training-free framework for constructing...

    arxiv.org/abs/2609.02160 · PDF

  42. 42

    Exact Limits of Random Projections for Preserving Geometry: Distance Recovery, Nearest-Neighbor Rankings, and Covariance Shape in Gaussian Models

    Piyush Sao

    cs.LG · cs.IT · math.NA · math.PR · math.ST

    The Johnson-Lindenstrauss (JL) lemma guarantees that a random projection of $n$ points to $m=O(\varepsilon^{-2}\log n)$ dimensions preserves pairwise squared distances within relative error $\varepsilon$ with high probability, and this dimension order is asymptotically optimal. In high dimensions, however, distances concentrate around a baseline while key geometric information lies in much smaller fluctuations. We show that the JL bound can...

    arxiv.org/abs/2609.02155 · PDF

  43. 43

    Online Non-Monotone DR-Submodular Maximization Matching the Offline $0.401$ Factor

    Vaneet Aggarwal, Yiyang Lu

    cs.LG · cs.AI · cs.CC · stat.ML

    We study online maximization of nonnegative, non-monotone DR-submodular functions over compact convex down-closed subsets of the $d$-dimensional unit cube. The best known constructive offline approximation factor is $0.401$ under the corresponding meta-solvability assumptions, whereas comparable adversarial online guarantees had remained at $1/e$. We show that this factor is also achievable online. In the post-decision full-information...

    arxiv.org/abs/2609.02145 · PDF

  44. 44

    Scalable Bayesian Optimization of Composite Functions for Image-Based Inverse Problems in Materials Characterization

    Dasol Yoon, Poompol Buathong, Chia-Hao Lee, Yujia Zhang, David A. Muller, Peter I. Frazier

    cs.LG · cond-mat.mtrl-sci

    Estimating physical parameters from scientific images is a common inverse problem in materials characterization that often relies on expensive physics-based simulations. In electron microscopy, specimen thickness and crystal mistilt are critical parameters that govern how electrons scatter through the sample, and therefore the accuracy of any atomic-scale structure recovered from it. They are commonly inferred by matching experimental...

    arxiv.org/abs/2609.02126 · PDF

  45. 45

    A Computational Comparison of Fourier Spectral Differentiation and Spatial Automatic Differentiation in Periodic Physics-Informed Neural Networks

    Xilai Liang, Zhao Zhang

    cs.LG

    Physics-informed neural networks (PINNs) commonly evaluate the spatial derivatives appearing in partial differential equation residuals using automatic differentiation (AD), whose computational and memory costs can become substantial when multiple or high-order derivatives are required. We perform a controlled comparison of spatial AD and Fourier spectral differentiation in periodic physical-space PINNs. Within each paired experiment, the...

    arxiv.org/abs/2609.02110 · PDF

  46. 46

    A Unified Rate-Distortion Perspective on Vector, Product, and Scalar Quantization

    Xianghong Fang, Wenlong Mou, Yuan Yuan, Dehan Kong, Tim G. J. Rudner

    cs.LG · cs.CV

    Discrete visual tokenization, predominantly driven by vector, scalar, and product quantization, lacks a unified conceptual framework for understanding quantization tradeoffs. In this paper, we propose a unified rate--distortion perspective on modern discrete visual tokenization. By viewing quantization as lossy compression, we characterize the nominal fixed-length coding rate through token count and codebook size, and quantization error as...

    arxiv.org/abs/2609.02107 · PDF

  47. 47

    Federated LoRA Adaptation of BiomedCLIP Across Four International Chest X-Ray Cohorts

    Sanjaya Poudel, Nirajan Kunwor, Manish Dhakal, Debesh Jha, Sunil Kumar Gaire

    cs.LG · cs.AI · cs.CV

    Federated learning (FL) lets institutions train a shared model without exchanging data, and Low-Rank Adaptation (LoRA) makes this practical at scale by communicating only compact low-rank updates. Biomedical imaging is a compelling setting for this combination: patient data are archived behind privacy regulations, and institutions differ widely in scanners, protocols, and compute. Such heterogeneity raises the question of how federated LoRA...

    arxiv.org/abs/2609.02101 · PDF

  48. 48

    Compositional Spectral Prompts for LLM-based Online Time Series Forecasting

    Seungyoon Choi, Hyunchul Kim, Jae-Gil Lee, Chanyoung Park

    cs.LG

    To address the sequential and evolving nature of time series, the Online Time Series Forecasting (OTSF) task has been extensively studied in multiple domains. Existing research focuses on adapting to non-stationary environments by employing memory buffer-based retrieval strategies. However, we observe that such frameworks struggle with long-term adaptation and fail to generalize to unseen patterns. To this end, we introduce CoSPOT, an...

    arxiv.org/abs/2609.02093 · PDF

  49. 49

    TC-Next: Zero-Shot Multimodal Cyclone Forecasting

    Zhe Wang, Sijie Chen, Yiming Luo, Daehyun Kim, Chien-Yi Chang

    cs.LG · cs.CV

    We present TropicalCycloneNext (TC-Next), a multimodal deep learning model that forecasts tropical cyclone track and intensity at $6$-$24$ h leads by leveraging a foundation model's forecast fields of atmospheric kinematic and thermodynamic fields and GridSat infrared satellite imagery. Trained only on GraphCast forecasts over the Western Pacific (WP), yet reliant only on generic atmospheric variables, TC-Next on GraphCast lowers track error...

    arxiv.org/abs/2609.02085 · PDF

  50. 50

    XMerge: Cross-Axis Selection and Reconstructive Layer Merging for LLM Depth Compression

    Jundong Hu, Shekar Ramachandran

    cs.LG · cs.CL

    Removing complete transformer layers preserves a standard serving architecture, but existing depth-compression methods can lose substantial quality, and the loss varies unpredictably across models. We introduce XMerge, a post-training method with two components. Cross-axis selection identifies a block with low relative-magnitude and angular hidden-state change, and local boundary reconstruction re-fits the adjacent surviving block to match...

    arxiv.org/abs/2609.02083 · PDF

  51. 51

    DynG-Diff: A State-Aware Dynamic Guidance Diffusion Framework for Probabilistic Time Series Forecasting

    Zhente Zhang, Zhengwei Ni, Wei Fan

    cs.LG

    Probabilistic multivariate time series (MTS) forecasting is crucial for modeling complex dynamical systems. However, existing diffusion-based methods rely on task-specific conditional paradigms that lack flexibility and struggle with inherent "information heterogeneity"--the significantly varying noise levels and evolutionary patterns across variables. To address this, we propose DynG-Diff, a variable-sensitive dynamic guidance diffusion...

    arxiv.org/abs/2609.02068 · PDF

  52. 52

    The Dynamics of Continuous Mixture Collapse in Language Models

    Ali Backour

    cs.LG · cs.CL

    LLMs latent-state reasoning methods replace discrete intermediate tokens with continuous states, such as weighted mixtures of token embeddings, to retain multiple possible reasoning directions rather than committing to one. Yet pretrained language models often fail to preserve these mixtures. We study why through a combination of theoretical analysis and controlled empirical investigations on a variety of models. We identify three...

    arxiv.org/abs/2609.02049 · PDF

  53. 53

    Act More, Decide Less: Skill-Guided Adaptive Action Chunking for Long-Horizon LLM Agents

    Yanting Yang, Can Jin, Jinman Zhao, Jiahao Wu, Yang Zhou, Zhepeng Wang, Zhendong Wang, Mu Zhou, Dimitris N. Metaxas

    cs.LG

    Large language model (LLM) agents for long-horizon interactive tasks typically follow a ReAct-style protocol, issuing one primitive action per LLM round. While this enables frequent replanning, it is inefficient for long-horizon tasks where many rounds are spent on routine action sequences. A natural alternative is to let the agent emit variable-length action chunks. However, naively training such policies with standard reinforcement learning...

    arxiv.org/abs/2609.02042 · PDF

  54. 54

    Source-Free Class Relearning: Diagnosing Forgetting in Class Unlearning

    Zahra Dehghani, Pablo Piantanida, Mohammadhadi Shateri

    cs.LG · cs.CV

    Class unlearning aims to remove a model's ability to recognize designated forget classes while preserving performance on retain classes. However, low forget accuracy after unlearning does not necessarily mean the class structure has been erased. Approximate unlearning methods can alter classifier decision boundaries while leaving recoverable structure in the representation. Prior work has shown that forget classes can be recovered, but...

    arxiv.org/abs/2609.02018 · PDF

  55. 55

    Train What You Deploy: Closing the MLP Reachability Gap in Low-Rank Clone Distillation

    Wenhui Chen, Zhifeng Li, Jie Zhou, Navan Preet Singh, Madalina Ciobanu, Chenghua Wang, Qingqing Mao, Ritankar Das

    cs.LG · cs.CL

    A compressed student has two shapes that need not agree: the weight it deploys at inference and the weight family its training can reach. We show that a state-of-the-art weight-inheritance distiller, Low-Rank Clone (LRC), deploys a full-width student MLP but ties training to a teacher-induced slice, leaving 62.5-81.4% of each deployed matrix's independent linear degrees of freedom unreachable-paid for at inference, never trainable. Our...

    arxiv.org/abs/2609.02006 · PDF

  56. 56

    CAHR-Net: Condition-Adaptive Hysteresis Reconstruction for Compact and Interpretable Magnetic Core Loss Modeling

    Chunye Gong, Cong Yao

    cs.LG

    Magnetic core loss originates in the hysteresis loop: the energy dissipated per excitation cycle equals the loop area, and frequency, temperature, and waveform shape set the loss by reshaping the loop geometry. Most existing models let these conditions act only on a terminal scalar - empirical equations fold them into fitted exponents, and data-driven predictors append them to encoded features - so no intermediate hysteresis representation...

    arxiv.org/abs/2609.01991 · PDF

  57. 57

    A Unified Particle Filter LSTM for Data-Driven Process Simulation

    Parvin Malekzadeh, Opher Baron, Dmitry Krass

    cs.LG · eess.SP

    Data-driven process simulation aims to generate realistic case trajectories from historical event logs without requiring an explicitly specified model of the underlying dynamics. Deep sequence models can capture complex temporal dependencies through next-activity probabilities and conditional time distributions. However, event logs provide only a partial view of the underlying process state, often recording activity completions without the...

    arxiv.org/abs/2609.01967 · PDF

  58. 58

    FlashKAN: B-Spline KANs via Truncated Power Form

    Naveen Mysore

    cs.LG · math.NA

    Kolmogorov-Arnold Networks (KANs) place learnable B-spline activations on network edges rather than fixed activations on nodes. The standard Cox-de Boor recursion evaluates these activations through k sequential passes for degree-k splines, consuming over 90% of forward-pass time. FlashKAN replaces this recursion with the truncated power form, a classical result from approximation theory that expresses each uniform cubic B-spline as five...

    arxiv.org/abs/2609.01956 · PDF

  59. 59

    Convergence Theory of Knowledge Distillation in Asynchronous P2P Gossip Learning Network

    Lucas Qingyang Fang, Tiyao Liu, Jinhao Jing, Zeji Li, Kaijie Chen, Harikrishna Kuttivelil, Katia Obraczka

    cs.LG · cs.AI

    Decentralized, serverless learning increasingly connects devices running different architectures, where the standard tool, decentralized SGD, is undefined as models with different parameter counts cannot be averaged. Knowledge distillation (KD) exchanges soft predictions rather than weights and sidesteps this obstacle, yet convergence theory for fully decentralized, asynchronous peer-to-peer (P2P) KD is lacking. We provide one, relocating...

    arxiv.org/abs/2609.01952 · 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.