cs.LG · 2026-07-21 · No. 60

Machine Learning, 2026-07-21.

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

    Causal Discovery on Irregular Time Series

    Martim Penim, Ricardo Ribeiro Pereira, Jacopo Bono, Hugo Ferreira, Mário A. T. Figueiredo, Pedro Bizarro

    cs.LG · stat.ME

    Causal discovery methods have shown strong performance in temporal systems, but they typically rely on regular and discrete lag structures, limiting their applicability to regularly sampled data. However, many real-world tasks require dealing with irregularly sampled streams of events, such as sensor streams, healthcare data, and financial transactions. In this work, we propose an extension of PCMCI+, a state-of-the-art method for causal...

    arxiv.org/abs/2607.18226 · PDF

  2. 02

    Three-Body Scattering for Generative Modeling

    Peng Sun, Zhenglin Cheng, Deyuan Liu, Jun Xie, Xinyi Shang, Tao Lin

    cs.LG · cs.CV

    Modern generative models typically rely on an adversarial critic, a prescribed noise-to-data path, or an autoregressive factorization. Instead, we show that a proper distributional energy can induce sample-level motion and provide direct regression supervision for a one-step generator. Three-Body Scattering Modeling (TBSM) for generation turns the energy distance into a constant-size per-projectile interaction: each projectile is attracted...

    arxiv.org/abs/2607.18198 · PDF

  3. 03

    FlashRT: Agent Harness for Guiding Agents to Deploy Real-Time Multimodal Applications

    Krish Agarwal, Zhuoming Chen, Yanyuan Qin, Zhenyu Gu, Atri Rudra, Beidi Chen

    cs.LG

    Real-time multimodal applications, including voice agents and interactive video generation, compose heterogeneous models into pipelines whose efficient deployment requires application-specific decisions about placement, streaming, and intra-model parallelism. Existing serving systems and auto-parallelism compilers commit to limited transformations and fixed workload assumptions, so achieving high performance on a new application requires...

    arxiv.org/abs/2607.18171 · PDF

  4. 04

    A Continual Validation, Updating, and Decision-Making Framework for Self-Adaptive Digital Twins via Robust Model Predictive Control: A Case Study in Additive Manufacturing

    Yi-Ping Chen, Ying-Kuan Tsai, Vispi Karkaria, Seul Lee, Daniel Apley, Wei Chen

    cs.LG · cs.AI · math.ST

    Digital Twins rely on surrogate models to mirror physical systems in real time, yet these models can degrade as operating conditions evolve, a phenomenon known as concept drift. Maintaining surrogate fidelity under drift, particularly when models must also capture aleatoric uncertainty, remains an open challenge. Existing adaptive frameworks lack principled mechanisms for detecting when updates are needed, for efficiently adapting models from...

    arxiv.org/abs/2607.18164 · PDF

  5. 05

    OR Else: A Differentiable Trust Region for Policy Optimization

    Chinmay Rane, Kanishka Tyagi, Michael Manry

    cs.LG · cs.AI

    PPO and the GRPO baseline studied here use clipped surrogate objectives whose favorable-direction saturation introduces an abrupt change in the scalar objective's derivative. We ask whether Output Reset (OR), a smooth one-sided saturation rule, offers a useful alternative for large language model post-training. PPO-OR and GRPO-OR replace the clipped policy term with an OR squared-margin loss in rollout-relative token log-ratio space; the...

    arxiv.org/abs/2607.18163 · PDF

  6. 06

    The Calibration Channel Determines the Bayes-Error Proxy: An Exact Law for Temperature-Induced Distortion

    Shreyas Pradeepkumar Khandale

    cs.LG

    The soft-label Bayes-error estimator beta(z) = E[min(z, 1-z)] of Ishida et al. estimates the irreducible error of a binary task directly from probability-valued labels. Recent work by Ushio et al. showed that this estimator is fragile when the probabilities are not the true posterior: even perfectly calibrated soft labels can yield a substantially inaccurate estimate, and they propose isotonic calibration as a consistent remedy. We complement...

    arxiv.org/abs/2607.18162 · PDF

  7. 07

    Differentiable Logic Gate Networks for Low-Latency EEG Classification on Edge Devices

    Shyamal Y. Dharia, Stephen D. Smith, Camilo E. Valderrama

    cs.LG · cs.AI

    Real-time EEG classification on edge devices is bottlenecked by the floating-point arithmetic of conventional neural networks. We investigated Differentiable Logic Gate Networks (Diff-Logic) as a hardware-native alternative that compiles models into pure Boolean circuits executable via bitwise CPU operations. Through rigorous iso-parameter experiments across four EEG datasets spanning two classification tasks, binary dementia detection and...

    arxiv.org/abs/2607.18149 · PDF

  8. 08

    Totally Positive Matrices and the Highest-Order Coefficients of the Characteristic Polynomial

    Tiago Closs, Leandro Farina

    cs.LG · math.RA

    We investigate the extent to which totally positive matrices can be distinguished through the highest-order coefficients of their characteristic polynomials. To identify the most informative coefficients, we also employed neural-network classifiers together with feature-attribution methods. Using datasets built from several structured totally positive families, including products of positive bidiagonal matrices, Vandermonde matrices, and...

    arxiv.org/abs/2607.18148 · PDF

  9. 09

    Do Language Models Dream of Binding Molecules? Benchmarking LLMs under Spatial Constraints

    Thomas MacDougall, Maksim Kuznetsov, Roman Schutski, Rim Shayakhmetov, Maxim Malkov, Vladimir Aladinskiy, Alex...

    cs.LG · cs.AI · cs.CL

    Structure-based drug design (SBDD) leverages the 3D structure of protein targets, often complemented by other spatial constraints, to generate candidate binding molecules. While diffusion models have dominated as a leading paradigm for high-quality 3D molecule generation, LLM-based methods are rapidly emerging in molecular design and have shown competitive performance in pocket-conditioned molecular generation. However, their ability to...

    arxiv.org/abs/2607.18144 · PDF

  10. 10

    Manifold-Constrained Hyper-Connections for Parameter-Efficient Finetuning

    Valentijn Oldenburg, Floris de Kam, Bente Zuijdam, Lieve Eberson, Nicky van Zutphen, Stef de Wildt, Ivo Verhoeven

    cs.LG

    Most parameter-efficient finetuning (PEFT) methods adapt weights or activations, thus leaving one of the key Transformer components unchanged: residual connections. This paper investigates Manifold-Constrained Hyper-Connections (mHC), a generalisation of residual connections, as a novel PEFT approach, wrapping frozen OLMo-2 backbones with learned residual routing modules. We find that mHC can finetune frozen Transformers, but that its role...

    arxiv.org/abs/2607.18130 · PDF

  11. 11

    LLM-as-a-Coach: Experiential Learning for Non-Verifiable Tasks

    Tianzhu Ye, Li Dong, Guanheng Chen, He Zhu, Xun Wu, Shaohan Huang, Furu Wei

    cs.LG · cs.CL

    Reinforcement learning (RL) on open-ended tasks compresses an LLM's rubric-based evaluation into a scalar reward, discarding rich textual feedback and conflating responses with distinct quality profiles. We propose Experiential Learning (EL), which repurposes the feedback model from an LLM-as-a-Judge into an LLM-as-a-Coach. The coach distills its assessment of each on-policy response into transferable experiential knowledge, which conditions...

    arxiv.org/abs/2607.18110 · PDF

  12. 12

    Empowering On-Device Model Adaptation with an Edge AI Inference Accelerator

    Mateusz Piechocki, Alessandro Capotondi, Marek Kraft

    cs.LG · cs.AR · cs.CV

    On-device model adaptation is essential to enable lifelong personalization on resource-constrained hardware, but compute, power, and memory limitations of such devices make end-to-end backpropagation impractical for modern deep neural networks. This work proposes a heterogeneous adaptation pipeline that repurposes a commercial edge AI inference accelerator, Hailo-8L, for frozen-backbone feature extraction during on-device training. The...

    arxiv.org/abs/2607.18101 · PDF

  13. 13

    The Label Complexity of Class-Conditional Coverage under Distribution Shift

    Weijia Han, Lisha Qu

    cs.LG · cs.CV

    Standard evaluation of many recognition systems contains distribution shift by construction, since benchmarks place disjoint conditions in the training and test splits. Under such a shift, split conformal prediction keeps marginal coverage near the nominal level while per-class coverage fails silently: on a real cross-subject skeleton benchmark, marginal coverage stays near ninety percent, the worst action class is covered about seventy...

    arxiv.org/abs/2607.18088 · PDF

  14. 14

    Enhancing Rubric-based RL via Self-Distillation

    Mingxuan Xia, Yuhang Yang, Chao Ye, Shuai Zhu, Shenzhi Yang, Guangcheng Zhu, Yuhang Zhang, Cheng Peng, Haobo Wang,...

    cs.LG · cs.AI

    Rubric-based RL has recently shown promise in improving LLMs on open-ended tasks. A widely recognized limitation of rubric-based RL is limited exploration: criteria that no rollout manages to satisfy (Unexplored Criteria, UC) receive no optimization signal. Recent methods address this by incorporating rubric information as external guidance during rollout, yet they introduce a train-inference mismatch: the policy is optimized on rollouts...

    arxiv.org/abs/2607.18082 · PDF

  15. 15

    SelectInfer: Selective Neuron Loading and Computation for On-Device LLMs

    Huzaifa Shaaban Kabakibo, Eric Schniedermeyer, Artem Burchanow, Lin Wang

    cs.LG · cs.AI

    Large Language Models (LLMs) have demonstrated remarkable capabilities across a range of Natural Language Processing (NLP) tasks, but their high computational and memory demands pose significant challenges for deployment on resource-constrained edge devices. Existing approaches to model compression and optimization often rely on coarse-grained pruning or quantization, which can compromise accuracy or require re-training and fine-tuning. In...

    arxiv.org/abs/2607.18081 · PDF

  16. 16

    Generalised Bellman recurrence and three dualities in sequential decision-making

    Fernando E. Rosas, David Hyland, Daniel Polani

    cs.LG · cs.AI

    What gives the Bellman equation its form? We show that the recursive properties of optimal value functions follow from three conditions: that the dynamics decomposes through sufficient statistics, that the return decomposes recursively, and that the aggregation of uncertainty is compatible with both. When all three conditions hold on a common state, the Bellman equation arises from their mutual consistency; when one fails, tractability can...

    arxiv.org/abs/2607.18077 · PDF

  17. 17

    Sobek: Streaming Equivariant Tensor Product Convolutions

    Vladimir Chorošajev, Cédric Bény

    cs.LG

    Equivariant graph neural networks repeatedly apply edge-conditioned tensor-product convolutions over graph edges. Conventional implementations materialize edge-specific weights, messages, and adjoints, causing tensor-product workspace and memory traffic to grow rapidly with graph size and operator width. This limits feasible workloads and can prevent larger problems from fully utilizing the GPU. We show that these edge-sized intermediates are...

    arxiv.org/abs/2607.18074 · PDF

  18. 18

    SGN: A Similarity-based Generative Network for Data Generation under Distribution Shift

    Jiaqi Zhu, Xincheng Chen, Yuncheng Wu, Zhaojing Luo, Beng Chin Ooi

    cs.LG · cs.AI

    Generative models trained on a source domain often produce samples that are poorly aligned with shifted target domains, limiting their effectiveness for target-domain data augmentation. Although target-specific adaptation can reduce this mismatch, it typically requires additional optimization and domain-specific parameters. We propose a Similarity-based Generative Network (SGN), a reusable framework that is trained once on labeled source data...

    arxiv.org/abs/2607.18072 · PDF

  19. 19

    SEE: Structure-aware Exploring \& Exploiting for Long-horizon GUI Agent Trajectory Synthesis

    Zhuohang Fan, Beichen Zhang, Yuanfa Li, Changqiao Wu, Wei Liu, Jian Luan, Weigang Zhang

    cs.LG

    Graphical User Interface (GUI) agents powered by vision-language models hold promise for automating real-world mobile tasks. However, progress is limited by the lack of high-coverage, long-horizon interaction trajectories collected from element-rich and rapidly evolving apps. Existing pipelines often rely on costly human demonstrations or on-policy framework, which tends to over-sample common flows while missing rare transitions and complex...

    arxiv.org/abs/2607.18046 · PDF

  20. 20

    Adaptive Mamba Neural Operators

    Zeyuan Song, Zheyu Jiang

    cs.LG · math.AP · math.NA

    Accurately solving partial differential equations (PDEs) on arbitrary geometries and a variety of meshes is an important task in science and engineering applications. In this paper, we propose Adaptive Mamba Neural Operators (AMO), which integrates reproducing kernels for state-space models (SSMs) rather than the kernel integral formulation of SSMs. This is achieved by constructing Takenaka-Malmquist systems for the PDEs. AMO offers new...

    arxiv.org/abs/2607.18043 · PDF

  21. 21

    L1 Augmented Attention as an Improved Vector Similarity Metric

    Kurt Godden

    cs.LG · cs.CL

    Scaled dot product attention conflates directional alignment and vector magnitude, limiting its effectiveness as a similarity metric in Transformer models. We introduce L1 augmented attention, a simple and computationally parallelizable modification that subtracts a learned, head specific L1 distance between queries and keys from the dot product score. This hybrid similarity captures complementary geometric information. Dot product rewards...

    arxiv.org/abs/2607.18027 · PDF

  22. 22

    FlashPDE: A Drop-in Fused Triton Operator Library for Neural PDE Solvers

    Peiyu Zang, Bosen Xie, Ruoxiang Xu, Yongqiang Cai

    cs.LG · cs.MS

    Physics-Informed Neural Networks (PINNs) solve PDEs by incorporating physical constraints into neural-network training, but large-scale problems are limited by automatic-differentiation memory overhead and inefficient execution of grid-based PDE operators. We present FlashPDE, a drop-in fused operator library for grid-based scientific machine learning. FlashPDE replaces fragmented PyTorch finite-difference execution with differentiable Triton...

    arxiv.org/abs/2607.18020 · PDF

  23. 23

    MADA-RL: Multi-Agent Debate-Aware Reinforcement Learning for Parameter-Efficient Reasoning in Compact Models

    Martino M. L. Pulici, Cuong Xuan Chu, Evgeny Kharlamov, Zifeng Ding, Volker Tresp, Yunpu Ma

    cs.LG · cs.AI · cs.CL · cs.MA

    Large language models achieve strong reasoning performance, but often at prohibitive training cost - a challenge that is especially acute for compact models ($\leq 4 \, \mathrm{B}$ parameters) trained under limited budgets. We introduce MADA-RL, a post-training framework that specializes compact models into generator and critic roles and trains them with a debate-aware learning signal, fine-tuning only a small subset of parameters via LoRA...

    arxiv.org/abs/2607.18006 · PDF

  24. 24

    fSRD: Fuzzy Spectral Region Decomposition -- Automated Multi Operator Koopman Representations via an Adaptive Spectral Learning Architecture

    Charles Bokor, Mark Cary, Denise Morrey, Fabrizio Bonatesta

    cs.LG

    Highly nonlinear chaotic dynamical systems remain difficult to model due to fundamental trade-offs between complexity, expressivity, and data efficiency. Modern machine learning methods achieve strong predictive performance but often rely on a-priori system knowledge or curated data with limited interpretability. Koopman operator theory offers a promising direction via linear representation in an infinite-dimensional observable space....

    arxiv.org/abs/2607.17990 · PDF

  25. 25

    Information-Based Exploration via Random Features for Reinforcement Learning

    Waris Radji, Odalric-Ambrym Maillard

    cs.LG

    Representation learning has enabled classical exploration strategies to be extended to deep Reinforcement Learning (RL), but often makes algorithms more complex and theoretical guarantees harder to establish. We introduce Random Feature Information Gain (RFIG), grounded in Bayesian kernel methods theory, which uses random Fourier features to approximate information gain and compute exploration bonuses in non-countable spaces. We provide error...

    arxiv.org/abs/2607.17981 · PDF

  26. 26

    Harness Engineering for LLM-Driven GPU Kernel Generation

    Yue Shui, Chenyu Ma, Hangfei Xu, Shengzhao Wen, Yanpeng Wang

    cs.LG · cs.AI

    Large language models (LLMs) can assist GPU kernel generation, but their practical effectiveness depends on whether generated code can be reliably constrained, validated, profiled, and selected. This paper presents a harness-centered system for LLM-driven GPU kernel optimization in the MLSys 2026 FlashInfer AI Kernel Generation Contest on NVIDIA Blackwell B200 GPUs. The system separates an evaluation harness from a profile-backed optimization...

    arxiv.org/abs/2607.17979 · PDF

  27. 27

    DiFA: Inference-Time Forward-Process Alignment for Diffusion Models

    Shigui Li, Delu Zeng

    cs.LG

    The prevailing inference framework for diffusion models formulates generation fundamentally as a problem of numerical integration. This perspective casts the model as an exact estimator, neglecting the inherent statistical uncertainty of the denoising process. In this work, we propose Forward-Process Aligned Diffusion prediction (\textbf{DiFA}), a training-free framework that reframes inference-time data prediction refinement as a sequential...

    arxiv.org/abs/2607.17972 · PDF

  28. 28

    Topological Signatures of Context-Level Reliability in TabPFN

    James Hu, Mahdi Ghelichi

    cs.LG · cs.AI · stat.ML

    TabPFN is a transformer-based foundation model for tabular prediction that performs inference without task-specific training by conditioning on a support set and query inputs. Despite its strong empirical performance, its internal behavior on structurally difficult tabular geometries remains poorly understood. We study this behavior using zigzag persistent homology, treating TabPFN layer representations as evolving point clouds. We construct...

    arxiv.org/abs/2607.17962 · PDF

  29. 29

    A Geometric Perspective on Stabilizing Value Conflict Resolution

    Saket Reddy, Andy Liu

    cs.LG · cs.AI

    Large Language Models (LLMs) often struggle to navigate value conflicts when trained with the compressed scalar rewards of Reinforcement Learning from Human Feedback (RLHF). To address this challenge, we investigate how chain-of-thought (CoT) reasoning can help improve performance in this domain. Geometrically, we show that CoT correlates with further smoothing the model's loss landscape in its sharpest direction, helping resolve the...

    arxiv.org/abs/2607.17946 · PDF

  30. 30

    The Art of Not Forgetting

    Ashmith Atmuri, Akshay Kumar, Yashaswini Rao Bhogarajula

    cs.LG · cs.AI

    We introduce CMP (Cognitive Memory Primitive), an architecture that represents inputs as sparse relational codes, stores them in a two-tier competitive memory, and learns entirely through local, gradient-free updates, with no backpropagation anywhere in the network. We use this architecture to test a specific hypothesis: that catastrophic forgetting, usually treated as a training-time defect to be patched with replay or regularization, is...

    arxiv.org/abs/2607.17944 · PDF

  31. 31

    Distributional Soft Bellman Operator under the Cramér Geometry

    Keru Wang, Yixin Deng, Yao Lyu, Stephen Redmond, Shengbo Eben Li

    cs.LG

    Distributional soft policy iteration (DSPI) provides an important framework for combining distributional reinforcement learning (DRL) with maximum-entropy control, in which the policy evaluation step is governed by a distributional soft Bellman operator acting on entropy-regularised returns. Theoretical analysis of such an evaluation step requires a probability metric under which Bellman updates can be controlled, typically by showing that...

    arxiv.org/abs/2607.17897 · PDF

  32. 32

    Mobius Learning: Cyclic Depth Folding in Transformers

    Tongtian Zhu

    cs.LG · cs.CL · cs.DC

    Transformer-based language models organize computation along an ordered depth axis, where shallow and deep blocks often develop distinct representational roles. We challenge the conventional view that these roles must remain tied to a block's position in the ordered sequence. We introduce Mobius Learning, a training architecture based on cyclic depth folding, in which different data streams follow cyclically shifted block orders. The same...

    arxiv.org/abs/2607.17843 · PDF

  33. 33

    Theoretical Foundations of $\max$@$k$ Reinforcement Learning

    Riccardo Poiani, Martino Bernasconi, Andrea Celli

    cs.LG

    Reinforcement Learning is a cornerstone technique for modern large reasoning models. Usually, for difficult tasks such as code generation and theorem proving, the agent is evaluated by generating $K$ responses rather than sampling a single response, and performance is then measured using a retry-aware metric such as $\max$@$k$. Despite their practical importance, the theoretical foundations of learning under such criteria remain limited. In...

    arxiv.org/abs/2607.17823 · PDF

  34. 34

    Phasor Attention: Mean Root Square Normalization for Phase Manifold Preservation

    Sungwoo Goo, Hwi-yeol Yun, Sangkeun Jung

    cs.LG · cs.AI

    While Root Mean Square Normalization has become the de facto standard for accelerating modern sequence models, its reliance on the quadratic accumulation of independent scalars ($\sum x^2$) inherently triggers outlier-induced numerical instability, gradient starvation, and anisotropic phase distortion. We introduce Mean Root Square Normalization (MRSNorm). By structurally pairing channels into 2D phasors, MRSNorm mathematically inverts the...

    arxiv.org/abs/2607.17822 · PDF

  35. 35

    The Concept of Representation in ML: Beyond Plato and Aristotle

    Gilad Landau, Aviv Keren

    cs.LG

    Representation is a central concept in modern machine learning, where it usually refers to internal encodings that support learning and generalization. As models scale and their capabilities become increasingly human-level, this representational language sometimes shifts from an engineering context into the more philosophically loaded domain of mental representation. We argue that this is the case for recent claims about the convergence of...

    arxiv.org/abs/2607.17800 · PDF

  36. 36

    Feature Attribution-Based Explainability Analysis of Deep Learning Models in Predictive Process Monitoring

    Kseniya Sahatova, Rafael Seidi Oyamada, Xuefei Lu, Johannes De Smedt

    cs.LG · cs.AI

    Predictive process monitoring supports the optimization and control of operational business processes by forecasting the future state or outcome of ongoing cases. While deep neural networks have achieved strong performance for these tasks by modeling sequential dependencies in event logs, their black-box nature limits trust and practical adoption. Feature attribution methods are often used to address this, but applying them directly poses a...

    arxiv.org/abs/2607.17783 · PDF

  37. 37

    FIFA World Cup 2026 as a Contamination-Free Benchmark for LLM Forecasting Agents: Four Models, a Bookmaker, and 104 Matches

    Jiacheng Ding, Cong Guo, Jason Xu

    cs.LG · cs.AI · cs.DB

    We introduce WC2026-Agents, a benchmark and dataset for evaluating large language models (LLMs) as autonomous forecasting agents on real, future events. For every one of the 104 matches of the 2026 FIFA World Cup, four frontier models -- Claude Opus 4.8, ChatGPT (GPT-5.5, high reasoning), Gemini 3.1 Pro, and Grok (Expert Mode) -- ran an identical search-act-reflect loop: gather evidence with a web tool, commit to a 1X2 (team-A win / draw /...

    arxiv.org/abs/2607.17765 · PDF

  38. 38

    Generalize and Guide: Decomposing Rewards for Few-Shot Inverse Reinforcement Learning

    Ziyi Liu, Grace Zhang

    cs.LG · cs.AI · cs.RO

    Inverse reinforcement learning (IRL) provides a powerful framework for learning from demonstrations. However, real-world tasks often exhibit substantial natural variations (e.g., picking up mugs with varying shapes), making it impractical to collect demonstrations that fully specify a new task under every possible scenario. In practice, while demonstrations for the target task are limited, it is often easier to obtain datasets of...

    arxiv.org/abs/2607.17760 · PDF

  39. 39

    Towards Reliable Zero-Shot Crowd Forecasting: Evaluating Time Series Foundation Models for Special Event Pedestrian Forecasting

    Ziteng Li, Yanan Xin, Tina Comes, Serge Hoogendoorn

    cs.LG

    Managing massive crowds during infrequent special events requires reliable real-time pedestrian-flow forecasting to ensure public safety and operational efficiency. However, supervised forecasting methods face limitations in these contexts due to scarce historical data, heterogeneous data distributions, and short in-event observation windows. To effectively support operational decision-making, forecasts should provide not only accurate point...

    arxiv.org/abs/2607.17758 · PDF

  40. 40

    MXSens: Sensitivity-Aware Mixed-Precision Quantization for Efficient LLM Inference

    Simla Burcu Harma, Danila Mishin, Zhengyuan Su, Ayan Chakraborty, Elizaveta Kostenok, Dongho Ha, Babak Falsafi,...

    cs.LG · cs.AI

    4-bit quantization enables efficient LLM inference, but suffers from significant accuracy degradation due to outliers. Prior work addresses this problem via data rotation or mixed-precision integer quantization, but often relies on software-managed scaling and frequent dequantization, incurring substantial overhead. Microscaling formats, such as MXINT, eliminate these inefficiencies by encoding scales in hardware, yet remain incompatible with...

    arxiv.org/abs/2607.17733 · PDF

  41. 41

    Planning with Transformers: Chain of Computation and Structured Context Windows

    Ehsan Futuhi, Nathan R. Sturtevant

    cs.LG

    Large Language Models (LLMs) have had a remarkable impact across many areas of machine learning. However, recent studies have shown that they struggle to reliably solve planning problems. At the same time, theoretical results have shown that transformers, the core architecture underlying modern LLMs, are Turing-complete. In this work, we investigate this apparent gap between the theoretical computational power of LLMs and their empirical...

    arxiv.org/abs/2607.17710 · PDF

  42. 42

    Uncovering Latent Reasoning Strategies in Language Models

    Awni Altabaa, John Lafferty

    cs.LG · cs.AI

    A language model $p_θ(y \mid x)$ trained on reasoning tasks learns to solve problems via multiple distinct strategies, yet these strategies are implicit and entangled within the model's response distribution. We study the problem of decomposing the response distribution of a given pretrained language model into a structured, strategy-conditioned representation. Specifically, we learn a latent-variable factorization $p_θ(y \mid x) \leadsto...

    arxiv.org/abs/2607.17674 · PDF

  43. 43

    Beyond Objective Expressivity: Geometry Preservation in Multimodal Contrastive Learning

    Tillmann Rheude, Roland Eils, Benjamin Wild

    cs.LG · cs.AI

    Contrastive learning is increasingly moving toward settings with three or more modalities instead of image-text pairs. Yet, extending models from pairwise to higher-order multimodal alignment can introduce optimization and representation challenges. We identify encoder Jacobian conditioning as a key factor in trimodal contrastive learning: poorly conditioned encoders exhibit collapsing or amplified singular-value spectra, leading to exploding...

    arxiv.org/abs/2607.17673 · PDF

  44. 44

    GeneSpeak-FP: Target and Compound Retrieval from Observed Cell-Level Perturbation Signatures

    Kseniia Vaniushkina, Jeongmin Lim, Jinyong Park

    cs.LG

    Large-scale single-cell perturbation atlases make it possible to ask an inverse question: given an observed transcriptional response, which annotated targets and compounds in a fixed library are most consistent with that response? We present \model, a Transformer retrieval model for this closed-library setting. Each input is a cell-level perturbation signature formed by contrasting one treated cell with a cell-line-specific mean DMSO...

    arxiv.org/abs/2607.17671 · PDF

  45. 45

    Selectivity Matters: Source Node Influence Pruning for Unsupervised Graph Domain Adaptation

    Ridong Han, Yawen Shen, Zhongnian Li, Tongfeng Sun, Xinzheng Xu, Abdulmotaleb El Saddik

    cs.LG · cs.AI

    Unsupervised Graph Domain Adaptation (UGDA) aims to facilitate knowledge transfer from a labeled source graph to an unlabeled target graph by mitigating cross-domain distribution shifts. Existing methods primarily focus on node-level feature alignment in latent spaces, relying on the implicit assumption that all source nodes contribute positively to the alignment. However, this assumption often fails because a node's semantic information is...

    arxiv.org/abs/2607.17668 · PDF

  46. 46

    TypiCore: A Hybrid Active Query Strategy for Class-Incremental Learning on Time Series

    Gabor Szucs, Samuel Jacsev, Marcell Nemeth, Davide Dalle Pezze, Gian Antonio Susto

    cs.LG · cs.AI

    Time series data play a pivotal role across numerous domains, including healthcare and manufacturing. In real-world environments, models must cope with distribution shifts over time, a challenge commonly addressed through Continual Learning (CL) techniques. However, existing CL methods face a critical limitation: real-world data streams are rarely fully labeled, making annotation cost a major practical constraint. This paper investigates...

    arxiv.org/abs/2607.17632 · PDF

  47. 47

    Can Transformers Really Do It All? On the Compatibility of Inductive Biases Across Tasks

    Damien Teney, Liangze Jiang, Hemanth Saratchandran, Simon Lucey

    cs.LG

    Transformers are remarkably versatile and their design is largely consistent across a variety of applications. But are they optimal for any given task or dataset? The answer may be key for pushing AI beyond merely scaling current designs. *Method.* We present a method to optimize a transformer architecture for a given dataset, which we use as a tool to study optimal task-specific inductive biases. This method replaces the most important...

    arxiv.org/abs/2607.17624 · PDF

  48. 48

    PoLoRA: A Preconditioned Orthogonalized LoRA Optimizer

    Nikhil Ghosh, Tetiana Parshakova, Robert M. Gower

    cs.LG · cs.CL · math.OC

    Low-rank adaptation (LoRA) makes finetuning large language models cheaper by adding to each weight matrix a trainable low-rank update parameterized as the product of two matrices. These matrices are usually trained with Adam, which treats them as a single flat vector of parameters and ignores both the matrix and product structure of LoRA. Applying a matrix-aware optimizer such as Muon to each factor does not consistently improve over Adam,...

    arxiv.org/abs/2607.17620 · PDF

  49. 49

    Optimizing the Preconditioner: A Black-box Online-to-Nonconvex Conversion with Static Regret Minimization Oracles

    Haichen Hu, David Simchi-Levi

    cs.LG · math.OC · stat.ML

    We study whether stochastic nonconvex optimization can be reduced to ordinary static regret minimization in online convex optimization in a black-box manner. For smooth nonconvex objectives, our reduction maintains a predictable gradient tracker, while a black-box online learner selects a preconditioner that determines how this tracker is transformed into the update direction. The learner receives linear convex losses and is evaluated against...

    arxiv.org/abs/2607.17607 · PDF

  50. 50

    Trustworthy Protein-Ligand Binding Affinity Prediction via Reliability-Aware Multi-Engine Fusion

    Yongchan Hong, Defu Cao, Wenjin Liu, Thomas Ku, Jordy Homing Lam, Emily Nguyen, Willie Neiswanger, Vsevolod Katritch, Yan Liu

    cs.LG · cs.AI

    Accurate protein-ligand binding affinity prediction is central to computational drug discovery, yet modern docking engines frequently disagree without indicating which prediction to trust. Consensus scoring and ensemble methods improve mean accuracy but treat all predictions identically without interpretable confidence measures or uncertainty decomposition, ignoring the chemical context of each protein-ligand pair. To address this limitation,...

    arxiv.org/abs/2607.17601 · PDF

  51. 51

    Concentration and Mean-Square Bounds for Contractive Stochastic Approximation: A Unified Elementary Approach

    Siddharth Chandak

    cs.LG · eess.SY · math.OC

    We establish mean-square and concentration bounds for stochastic approximation (SA) with arbitrary norm contractive mappings, under a multiplicative noise model where the noise may scale affinely with the norm of the iterates, and the iterates are potentially unbounded. These settings arise in reinforcement learning, where operators are often contractive in the $\ell_\infty$ norm and the noise scales with the iterates. To address the...

    arxiv.org/abs/2607.17595 · PDF

  52. 52

    ANNLib: A Development Framework for Efficient Approximate Nearest Neighbor Search

    Zheqi Shen, Jingbo Su, Zijin Wan, Yan Gu, Yihan Sun

    cs.LG · cs.IR

    Approximate Nearest Neighbor Search (ANNS) plays a pivotal role in modern deep learning pipelines. Recently, many ANNS systems have been proposed to either provide broad functionality or reach high performance. However, it is yet difficult to achieve both with minimal programming efforts. We propose ANNLib to address the gap. ANNLib is a library that provides a programming framework for achieving high performance and flexible functionality in...

    arxiv.org/abs/2607.17582 · PDF

  53. 53

    AGG: Jacobian-Aggregated Group Gradient for Efficient GRPO Training of Diffusion Models

    Ruiyi Ding, Jie Li, He Kang, Ziyan Liu, Chengru Song, Yuan chen

    cs.LG · cs.CV · eess.SY

    Group Relative Policy Optimization (GRPO) is a powerful reinforcement learning algorithm for aligning generative models with human preferences. While successful in large language models~\cite{shao2024deepseekmathpushinglimitsmathematical}, its extension to diffusion and flow matching models introduces a severe computational bottleneck: gradients must be back-propagated through the high-capacity DiT backbone at \emph{every} timestep of the...

    arxiv.org/abs/2607.17572 · PDF

  54. 54

    A Weisfeiler-Leman Characterization of Global-Attention Graph Transformers for Mixed-Integer Linear Programs

    Md Abrar Jahin, Craig A. Knoblock, Jay Pujara

    cs.LG

    Graph foundation models (GFMs) with global attention are increasingly used to represent mixed-integer linear programs (MILPs), aiming to capture structure beyond the locality of standard graph neural networks. We study their expressive power through graph isomorphism testing, asking which MILP instances they map to identical representations. We prove that a broad class of hierarchical graph transformers combining global linear attention,...

    arxiv.org/abs/2607.17570 · PDF

  55. 55

    CoCurve: Cross-Module Co-Pruning Curvature for Training-Free Structured LLM Pruning

    Zhiren Gong, Zihao Zeng, Zijie Wang, Tiantong Wang, Chau Yuen, Wei Yang Bryan Lim

    cs.LG · cs.AI

    Structured pruning compresses large language models (LLMs) by removing whole computational units, such as attention heads and feed-forward (FFN) channel groups. Most training-free methods, however, rank these units independently, implicitly treating the loss from pruning a set as the sum of its individual losses. This view fails for Transformers, whose sublayers are coupled through a shared residual stream. Two individually weak units can...

    arxiv.org/abs/2607.17568 · PDF

  56. 56

    Volatility-Aware Extreme Event Detection in High-Frequency Financial Markets

    Maorufa Zaman, Haris Md Sahed

    cs.LG

    Predicting extreme price movements in high-frequency financial markets is a challenging task due to non-stationarity, heavy-tailed return distributions, and severe class imbalance. In particular, rare but impactful events are often difficult to detect using conventional modeling approaches, which typically treat extreme movements as isolated observations. This study proposes a volatility-aware approach for extreme event detection using...

    arxiv.org/abs/2607.17555 · PDF

  57. 57

    Program Synthesis for Simulation-Based Inference: Joint Model Selection and Parameter Estimation

    Siddharth Mishra-Sharma

    cs.LG · stat.ML

    Neural simulation-based inference enables parameter estimation for complex models, but typically requires the user to specify a simulator encoding a fixed model structure. We present a framework for joint model selection and parameter estimation that combines large language models for program synthesis with neural simulation-based inference. Given a natural language description of the system and data under investigation, an LLM proposes...

    arxiv.org/abs/2607.17540 · PDF

  58. 58

    FailureAtlas: A Taxonomy of Failure Modes in Multi-Provider LLM Serving Infrastructure

    Vishal Pandey, Gopal Singh

    cs.LG · cs.SE

    Multi-provider LLM gateways reverse proxies that route, load-balance, and rate-limit requests across foundation-model APIs have become critical production infrastructure. Yet the failure modes specific to this architectural layer remain undocumented, scattered across issue trackers and post-mortems with no unifying framework. We introduce \fa{}, a two-axis taxonomy that classifies failures by their \emph{origin layer} (Network/Transport,...

    arxiv.org/abs/2607.17525 · PDF

  59. 59

    One-step lowest-variance selection in a Gaussian random-field model motivated by masked diffusion: Total correlation and a square root collision threshold

    Linjun Li

    cs.LG · cs.AI · cs.IT · math.PR

    Motivated by confidence-guided parallel unmasking in masked discrete diffusion, we study a single selection step in a stylized Gaussian random-field model. A locally dependent nonnegative score field represents position wise uncertainty, and the scheduler selects the K positions with the smallest scores. Dependence among the selected positions is measured through a distance-dependent Gaussian correlation model. This separation provides a...

    arxiv.org/abs/2607.17522 · PDF

  60. 60

    After the Euclidean Highway: Hyperbolic Expert AI as the Next Innovation

    Kwan Soo Shin, In Seok Kang, Munho Lee

    cs.LG · cs.AI · cs.CL

    Expert domains are trees; the Euclidean transformer is not, diluting parent-child structure exponentially at depth. The hyperbolic turn left one question unasked: not how much of a network to curve, but where curvature may touch the gradient. Placement is a law, not a knob: the same geometry on a trainable adapter collapses training (seventeen training collapses, ~220 GPU-hours), yet at the loss layer alone it trains without one -- this is...

    arxiv.org/abs/2607.17513 · PDF

  61. 61

    Lightweight Wrappers for Adapting Time Series Foundation Models to Regional Drought Forecasting

    Wentao Gao, Jiuyong Li, Lin Liu, Thuc Duy Le, Jixue Liu, Yanchang Zhao, Yun Chen

    cs.LG

    Large \emph{Time Series Foundation Models} (TSFMs) demonstrate strong zero-shot forecasting capabilities across diverse domains. However, their application to regional climate forecasting faces practical challenges: model weights are often proprietary, local training records are limited, and computational budgets are constrained, making traditional fine-tuning approaches infeasible. To address these constraints, we introduce a lightweight,...

    arxiv.org/abs/2607.17511 · PDF

  62. 62

    Retrieval-Augmented Interpretable Learning: Towards Task-Specific Zero-Shot Models in Healthcare

    Sazan Mahbub, Caleb Ellington, Zhiyuan Li, Yixin Yang, Souvik Kundu, Ben Lengerich, Eric P. Xing

    cs.LG · cs.AI

    We introduce Retrieval-Augmented Interpretable Learning (RAIL), a probabilistic meta-learning framework for zero-shot generation of task-specific interpretable models that synthesizes coefficient-space structure from natural-language task descriptions and a memory of previously learned task-specific predictors. RAIL retrieves related source tasks, transfers structure through coefficient space, and generates a new predictor in the original...

    arxiv.org/abs/2607.17508 · PDF

  63. 63

    Residual-Guided Multi-Resolution Refinement of Foundation Models: A Case Study in Drought Forecasting

    Wentao Gao, Jiuyong Li, Lin Liu, Thuc Duy Le, Jixue Liu, Yanchang Zhao, Yun Chen

    cs.LG

    Regional climate prediction presents unique challenges for time series foundation models, which typically process temporal patterns through single-pass inference. Expert climatologists, in contrast, employ multi-scale temporal analysis and iterative refinement based on systematic error diagnosis. We present RGMR (Residual-Guided Multi-Resolution Refinement), an inference-time framework that adapts pre-trained foundation models to perform...

    arxiv.org/abs/2607.17507 · PDF

  64. 64

    Calibrated Alzheimer's Conversion Risk in Mild Cognitive Impairment: Persistent Homology of Clinical Trajectories with Conformal Guarantees

    Navin Bondade

    cs.LG · stat.ML

    Background. Predicting conversion from mild cognitive impairment (MCI) to Alzheimer's disease (AD) is central to trial enrichment and care planning, yet existing models provide no individual-level uncertainty estimates and rarely include transparent leakage audits. We introduce the first application of persistent homology to longitudinal clinical trajectory point clouds for this task, and the first split-conformal individual risk guarantee...

    arxiv.org/abs/2607.17442 · 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.