cs.LG · 2026-07-01 · No. 40

Machine Learning, 2026-07-01.

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

01 — The papers

55 entries
  1. 01

    QVal: Cheaply Evaluating Dense Supervision Signals for Long-Horizon LLM Agents

    Sergio Hernández-Gutiérrez, Matteo Merler, Ilze Amanda Auzina, Joschka Strüber, Ameya Prabhu, Matthias Bethge

    cs.LG · cs.AI · cs.CL

    LLM agents increasingly act over long horizons, where a single trajectory can contain hundreds or thousands of actions. In these settings, outcome-only rewards provide too sparse guidance, failing to inform the model about the goodness of intermediate actions. Dense supervision methods aim to solve this problem by scoring intermediate steps, from intrinsic confidence to self-distillation and embedding similarities. However, it is common...

    arxiv.org/abs/2606.32034 · PDF

  2. 02

    AdaJEPA: An Adaptive Latent World Model

    Ying Wang, Oumayma Bounou, Yann LeCun, Mengye Ren

    cs.LG · cs.AI

    Latent world models enable planning from high-dimensional observations by predicting future states in a compact latent space. However, these models are typically kept frozen at test time: when their predictions become inaccurate, planning can fail, especially under test-time distribution shift. To address this, we propose AdaJEPA, an adaptive latent world model that performs test-time adaptation within the closed loop of model predictive...

    arxiv.org/abs/2606.32026 · PDF

  3. 03

    SemRF: A Semantic Reference Frame for Residual-Stream Dynamics in Language Models

    Jian Gu, Aldeida Aleti, Chunyang Chen, Hongyu Zhang

    cs.LG · cs.CL

    Residual-stream analysis asks how language-model computation evolves across depth, but intermediate decoding requires comparable readout coordinates across layers. If embedding anchors and unembedding readout disagree on the chosen span, apparent motion may reflect measurement drift rather than computation. We introduce \emph{Semantic Reference Frames} (SemRF), an anchor-based formalism separating semantic measurement from residual dynamics....

    arxiv.org/abs/2606.32022 · PDF

  4. 04

    TRIAGE: Role-Typed Credit Assignment for Agentic Reinforcement Learning

    Yuanda Xu, Zhengze Zhou, Hejian Sang, Xiaomin Li, Jiaxin Zhang, Xinchen Du, Zhipeng Wang, Alborz Geramifard

    cs.LG · cs.AI

    Agentic reinforcement learning requires assigning credit to environment-facing actions such as searches, clicks, edits, navigation commands, and object interactions. Standard GRPO uses the final verifier outcome as a uniform advantage over all action tokens. This outcome signal is useful but structurally incomplete: it punishes useful exploration in failed rollouts and reinforces redundant or regressive actions in successful rollouts. We...

    arxiv.org/abs/2606.32017 · PDF

  5. 05

    FedLAB: Traceable Semantic Codebooks for Federated Multimodal Graph Foundation Learning

    Zekai Chen, Kairui Yang, Xuaner Chen, Xunkai Li, Xun Wu, Rong-Hua Li, Guoren Wang

    cs.LG

    Multimodal graph foundation models aim to learn reusable knowledge from graphs enriched with text, images, attributes, and relational topology, thereby supporting diverse graph-centric and modality-centric tasks. In practice, however, such multimodal graphs are often distributed across decentralized clients, where raw contents and local structures cannot be centrally shared due to privacy constraints. This motivates federated multimodal graph...

    arxiv.org/abs/2606.32016 · PDF

  6. 06

    CoMet: Context and Multiplicity Decomposition for Multimodal Uncertainty Estimation

    Sanghyuk Chun, William Yang, Amaya Dharmasiri, Olga Russakovsky

    cs.LG · cs.CV

    Uncertainty estimation has been a long-standing challenge in AI models; it amounts to "knowing what you don't know," and metacognition is notoriously difficult even for humans (cf. the Dunning-Kruger effect). Although it is still far from solved even in simpler classification systems, tackling it in multimodal large language models (MLLMs) is becoming increasingly important. Within MLLMs, uncertainty can stem from any of the diverse sources...

    arxiv.org/abs/2606.32012 · PDF

  7. 07

    Surrogate Fidelity: When Can Open LLMs Explain Closed Ones?

    Philippe Chlenski, Zachariah Carmichael, Ayush Warikoo, Chia-Tse Shao, Yingxiao Ye, Aobo Yang, Vivek Miglani, Nehal Bandi

    cs.LG

    Mechanistic interpretability (MI) requires full access to model internals, yet the APIs for most widely deployed language models at best expose log-probabilities over output tokens. This creates a surrogate problem: when do measurements made on open models allow us to make claims about a closed model? We evaluate surrogate fidelity at the prediction, attribution, and representation levels. For binary classification tasks, log-odds provide an...

    arxiv.org/abs/2606.32008 · PDF

  8. 08

    Radial Suppression Accelerates Algorithmic Generalization: A Geometric Analysis of Delayed Generalization

    Srijan Tiwari, Aditya Chauhan, Manjot Singh

    cs.LG · cs.AI

    Why do neural networks memorize algorithmic training data long before they generalize? We present a geometric case study demonstrating that, on tasks where generalization requires discovering structured low-dimensional circuits, the memorization-generalization delay is driven by radial inflation of hidden representations under cross-entropy optimization. We formalize a radial-angular decomposition of activation-space dynamics and derive three...

    arxiv.org/abs/2606.32000 · PDF

  9. 09

    Amplifying Membership Signal Through Chained Regeneration

    Wojciech Łapacz, Stanisław Pawlak

    cs.LG · cs.AI

    The tendency of large generative models to memorize training data makes sample verification critical for privacy auditing and copyright enforcement. Current membership (MIA) and dataset inference (DI) attacks often rely on one-shot generations, which yield weak signals and limited sensitivity across modalities. Inspired by Model Autophagy Disorder (MAD), we introduce MADreMIA, a model-agnostic framework that enhances white-, gray-, and...

    arxiv.org/abs/2606.31991 · PDF

  10. 10

    Signed-Permutation Coordinate Transport for RMSNorm Transformers

    John Sweeney

    cs.LG · cs.CL · stat.ML

    Modern LLM workflows move coordinate-indexed objects across checkpoints: steering vectors, sparse autoencoders, top-$k$ neuron sets, attribution lists, and merge alignments. This is only well posed after fixing the model's residual-stream gauge, which we show is architecture-dependent: LayerNorm residual charts have permutation gauge $S_d$ (up to a global sign flip), while RMSNorm charts with generic per-channel gain have signed-permutation...

    arxiv.org/abs/2606.31963 · PDF

  11. 11

    Making Sense of Touch from the Child's View for Contrastive Learning

    Max Whitton, Zecheng Wang, Puchen Liu, Quang Tuan Truong, Shengao Wang, Manaswi Yadamreddy, Oktay Ozel, Visista...

    cs.LG

    Is the sense of touch a mechanism for human babies' learning of visual concepts? If so, can we quantify its importance, and to what extent do babies rely on their sense of touch for visual learning? To approach these questions in a principled way, we propose a structured coding system for baby-centric touch events, yielding a dataset of 264k two-second clips of touch events coded according to this system. Using this dataset, we pretrain...

    arxiv.org/abs/2606.31943 · PDF

  12. 12

    Interface-Aware Neural Newton Preconditioning for Robust Cohesive Zone Model Simulations

    Zhangyong Liang, Huanhuan Gao

    cs.LG · math.NA

    Cohesive Zone Models (CZMs) are widely used to simulate interface fracture, delamination, adhesive failure, and fiber--matrix debonding in aerospace composite structures. In implicit quasi-static finite element analyses, cohesive softening may introduce negative interface tangents, solution jumps, and Newton-basin mismatch, so the previous converged state can become a poor initial guess for the next increment. This may lead to stagnation,...

    arxiv.org/abs/2606.31921 · PDF

  13. 13

    Sequential RC-TGAN: Generating Relational Time Series with Spectral Envelope Loss

    Mohamed Gueye, Yazid Attabi, Manuel Morales, Maxime Dumas

    cs.LG

    The generation of synthetic relational databases often involves modeling complex temporal dynamics, such as transaction logs or event sequences. A significant challenge in this domain is the handling of categorical time series (e.g., status codes), where standard encoding methods like one-hot encoding fail to capture intrinsic frequency-domain features such as seasonality and cyclicity. In this paper, we introduce Sequential RC-TGAN (Seq....

    arxiv.org/abs/2606.31904 · PDF

  14. 14

    Review Residuals: Update-Conditioned Residual Gating for Transformers

    Kyle Kramer

    cs.LG · cs.CL

    Residual connections add every sublayer's proposed update with a fixed coefficient of one; the network never evaluates whether an update is reliable before committing it. Drawing on the human-factors principle of independent verification, we introduce Review Residuals, which scale each update by a learned, input-dependent gate conditioned on both the current state and the proposed update: h_l = h_{l-1} + r_l * u_l with r_l =...

    arxiv.org/abs/2606.31859 · PDF

  15. 15

    Low-dimensional topology of deep neural networks

    Junyu Ren, Lek-Heng Lim

    cs.LG · math.GT

    We study layered models, including feedforward networks, ResNets, and transformers, by limiting each layer to a width of $d = 3$, i.e., $\mathbb{R}^3$ as representation space. This allows us to track how a neural network changes low-dimensional topological invariants through its layers. Just about any topological structure may be simplified or even trivialized by simply increasing dimension; e.g., any knot is equivalent to an unknot in...

    arxiv.org/abs/2606.31856 · PDF

  16. 16

    Geometry-Preserving Orthonormal Initialization for Low-Rank Adaptation in RLVR

    Ruijia Zhang, Jiacheng Zhu, Hanqing Zhu, Laixi Shi

    cs.LG · cs.AI

    Low-rank adaptation (LoRA) and its variants enable parameter-efficient fine-tuning of large language models under the supervised fine-tuning (SFT) paradigm. However, their efficacy and behavior under Reinforcement learning with verifiable rewards (RLVR) are less well understood. In particular, two structurally initialized LoRA variants, PiSSA and MiLoRA, which outperform standard LoRA under SFT, can underperform standard LoRA under RLVR and...

    arxiv.org/abs/2606.31813 · PDF

  17. 17

    Relational and Sequential Conformal Inference for Energy Time Series over Graphs via Foundation Models

    Keivan Faghih Niresi, Alice Cicirello, Olga Fink

    cs.LG · stat.ML

    Accurate energy demand forecasting is essential for the reliable operation and planning of modern sustainable energy systems. Spatial-temporal graph neural networks (STGNNs) have recently achieved strong performance in point forecasting by jointly modeling temporal dynamics and relational dependencies across interconnected energy nodes. However, in real-world energy systems, accurate point forecasts alone are insufficient, as operators also...

    arxiv.org/abs/2606.31804 · PDF

  18. 18

    Bridging the Gap Between Latent and Explicit Reasoning with Looped Transformers

    Ying Fan, Anej Svete, Kangwook Lee

    cs.LG · cs.CL

    Language models typically reason via explicit chain-of-thought (CoT), generating intermediate steps token-by-token. Latent CoT offers an alternative: it performs multi-step reasoning in the model's hidden states, replacing decoded tokens with continuous representations for greater efficiency. However, existing latent CoT methods underperform explicit CoT beyond 1B parameters, and the gap widens with scale. Looped, or recurrent-depth,...

    arxiv.org/abs/2606.31779 · PDF

  19. 19

    Policy Optimization Achieves Data-Dependent Regret Bounds in MDPs with Unknown Transitions

    Mingyi Li, Taira Tsuchiya, Kenji Yamanishi

    cs.LG · stat.ML

    We study policy optimization for online episodic tabular Markov decision processes with unknown transition kernels, aiming for best-of-both-worlds guarantees together with data-dependent regret bounds. Recent work (Dann et al., 2023; Li et al., 2026) has shown that policy optimization can adapt to both adversarial and stochastic losses with first-order, second-order, and path-length bounds, but only under known transitions, leaving open...

    arxiv.org/abs/2606.31769 · PDF

  20. 20

    Addressing Over-Refusal in LLMs with Competing Rewards

    Taeyoun Kim, Aviral Kumar

    cs.LG

    Safety training on language models often induces over-refusal: improved safety on harmful prompts at the cost of increased refusal on harmless ones. Though this trade-off can be mitigated by training models with reinforcement learning (RL) to reason before answering, it does not remove the underlying problem that reasoning can often be a "rubber stamp" for a predetermined response. In this paper, we address the safety-refusal trade-off by...

    arxiv.org/abs/2606.31748 · PDF

  21. 21

    FedXDS: Leveraging Model Attribution Methods to counteract Data Heterogeneity in Federated Learning

    Maximilian Andreas Hoefler, Karsten Mueller, Wojciech Samek

    cs.LG · cs.AI

    Explainable AI (XAI) methods have demonstrated significant success in recent years at identifying relevant features in input data that drive deep learning model decisions, enhancing interpretability for users. However, the potential of XAI beyond providing model transparency has remained largely unexplored in adjacent machine learning domains. In this paper, we show for the first time how XAI can be utilized in the context of federated...

    arxiv.org/abs/2606.31742 · PDF

  22. 22

    Nonlinearity-Aware LoRA: Structured Gate Adaptation under Low-Rank Constraints

    Shuai Yuan, Sudong Cai, Bingzhi Chen, Shuyuan Zheng, Chuan Xiao, Makoto Onizuka, Rui Mao

    cs.LG

    Low-rank adaptation (LoRA) is commonly viewed as an update-space approximation to full fine-tuning, yet this view is incomplete for self-gated Transformer feed-forward networks. In gated FFNs, a low-rank residual can change not only projected features but also the nonlinear selection weights that determine which channels contribute to the output. We formalize this effect as selection misalignment and connect it to the local effective...

    arxiv.org/abs/2606.31717 · PDF

  23. 23

    Diffusing Blame: Task-Dependent Credit Assignment in Biologically Plausible Dual-Stream Networks

    Yutaro Yamada, Luca Grillotti, Rujikorn Charakorn, Sebastian Risi, David Ha, Robert Tjarko Lange

    cs.LG · cs.NE

    Biological neural circuits obey Dale's principle: each neuron's synapses are uniformly excitatory or inhibitory. Artificial networks that respect this constraint must coordinate separate excitatory and inhibitory populations, fundamentally changing how credit is assigned during learning. Several biologically plausible learning rules avoid backpropagation's weight transport requirement, but it has been difficult to achieve strong performance...

    arxiv.org/abs/2606.31700 · PDF

  24. 24

    When to Truncate a Feature Ranking: A Residual-Overlap Stopping Rule for Subset Selection

    Jesus S. Aguilar-Ruiz

    cs.LG · cs.AI

    Feature rankings are widely used in supervised feature selection because they are simple, scalable and easy to interpret. Variables are first ranked by a relevance score, and a subset is then obtained by retaining the top-ranked variables. Although the first stage has been extensively studied, the second is often governed by an arbitrary cardinality, an empirical threshold or cross-validation, without a direct interpretation. This raises a...

    arxiv.org/abs/2606.31686 · PDF

  25. 25

    Improving Certified Robustness via Adversarial Distillation

    Matteo Melis, Jesus Martinez Del Rincon, Vishal Sharma

    cs.LG · cs.AI

    Certified training aims to produce models whose predictions can be formally verified against adversarial perturbations, typically by optimising upper bounds on the worst-case loss over an allowed perturbation set. For neural networks, certified training methods based purely on tight relaxation bounds produce networks that are amenable to certification, but sacrifice standard accuracy. Conversely, adversarial training often yields stronger...

    arxiv.org/abs/2606.31653 · PDF

  26. 26

    ECHO: Prune to act, trace to learn with selective turn memory in agentic RL

    Zijun Xie, Binbin Zheng, Enlei Gong, Jihua Liu, Yuyang You, Lingfeng Liu, Jiayao Tang, Guanqun Zhao, Aoqi Hu, Zeyu Chen

    cs.LG · cs.AI

    Long-horizon language agents must repeatedly interact with tools, accumulate evidence, and make decisions under bounded context windows. Existing context-management methods make such rollouts feasible by truncating distant history, folding past turns into summaries, or selecting compact memory states. However, these breakthroughs introduce two coupled limitations. First, as the number of turns grows, historical observations are progressively...

    arxiv.org/abs/2606.31650 · PDF

  27. 27

    Calibration, Not Compilation: Detecting and Repairing Misspecified Probabilistic Programs Written by Language Models

    Jian Xu, Delu Zeng, John Paisley, Qibin Zhao

    cs.LG

    Language models increasingly write probabilistic programs (in NumPyro, Stan, or Pyro), but a program that compiles, runs, and passes every unit test can still be \emph{statistically} wrong -- a Gaussian likelihood for heavy-tailed data, a Poisson for over-dispersed counts, an invalid prior support, or a pathological parameterization. The right verifier is therefore not a test suite but the Bayesian workflow itself: posterior predictive...

    arxiv.org/abs/2606.31630 · PDF

  28. 28

    Evil Spectra: How Optimisers can Amplify or Suppress Emergent Misalignment

    Jason R. Brown, Patrick Leask, Lev McKinney

    cs.LG · cs.AI

    Emergent misalignment (EM) is a recently discovered phenomenon in LLMs where fine-tuning on a narrow misaligned task, such as writing insecure code, leads to broadly misaligned behaviour on unrelated prompts. Previous work has noted that the severity of EM is highly sensitive to training choices; however, we still lack a systematic characterisation of this sensitivity. We perform a sweep over several Qwen3 models, optimisers, datasets, and...

    arxiv.org/abs/2606.31591 · PDF

  29. 29

    Robustness of neural networks to random noise perturbations of their inputs

    Mark Levene, Martyn Harris

    cs.LG · cs.NE

    We investigate the problem of the robustness of a trained neural network to the perturbation of its input values. More specifically, we examine the interplay between the accuracy of the network, as measured by the mean squared error, and robustness. Accordingly, we present a robustness measure, which, with high probability, suggests an upper bound on the mean squared error of the network, with respect to an input data set, for a given...

    arxiv.org/abs/2606.31581 · PDF

  30. 30

    Introduction to Stochastic Differential Equations for Generative Machine Learning: A Variational Perspective

    Ole Winther, Paul Jeha, Sander Dieleman, Andriy Mnih, Manfred Opper, Andrea Dittadi

    cs.LG

    The use of ordinary and stochastic differential equations has led to substantial progress in generative machine learning with applications to, for example, image, video and biomolecule generation. This paper provides a self-contained and informal introduction to the differential equations, the probabilistic framework for using them in generative modeling and the Fokker--Planck equation that governs the temporal evolution of the marginal...

    arxiv.org/abs/2606.31576 · PDF

  31. 31

    Beyond the Expressivity-Trainability Paradox: A Dynamical Lie Algebra Perspective on Navigating Barren Plateaus in Quantum Machine Learning

    Kung-Ming Lan

    cs.LG · quant-ph

    As Quantum Machine Learning (QML) transitions toward practical implementation, the field faces a critical architectural bottleneck that challenges the fundamental assumptions of classical statistical learning theory. In classical deep learning, increasing model capacity typically risks overfitting. However, this study advances a counter-intuitive paradigm: unstructured contemporary QML architectures suffer from a profound state of quantum...

    arxiv.org/abs/2606.31536 · PDF

  32. 32

    On the Convergence of Self-Improving Online LLM Alignment

    Xudong Wu, Pangpang Liu, Vaneet Aggarwal, Jiayu Chen

    cs.LG · cs.AI · stat.ML

    The Self-Improving Alignment (SAIL) algorithm addresses distribution shift by reducing a bilevel formulation of the problem to an efficient, single-level method. Empirically, SAIL has demonstrated strong performance on this task. However, a formal analysis of its convergence properties has been lacking. We identify a key theoretical challenge: the standard SAIL objective function is not guaranteed to be strongly concave due to unfavorable...

    arxiv.org/abs/2606.31524 · PDF

  33. 33

    RaBitQCache: Rotated Binary Quantization for KVCache in Long Context LLM Inference

    Wenhao Li, Jinhao Dong, Hailin Zhang, Wenhang Shi, Wei Lu, Xiaoyong Du

    cs.LG · cs.CL

    Long-context Large Language Model inference is severely bottlenecked by the massive Key-Value (KV) cache, yet existing sparse attention methods often suffer from static fixed-budget (Top-k) retrieval or rely on proxy scores that are computationally expensive and biased. To address these limitations, we propose RaBitQCache, a novel sparse attention framework that utilizes randomized rotated binary quantization and high-throughput binary-INT4...

    arxiv.org/abs/2606.31519 · PDF

  34. 34

    Fork-Think with Confidence

    Zena Al-Khalili, Rafi Hakim, Dietrich Klakow, Ji-Ung Lee

    cs.LG · cs.CL

    Parallel thinking has enjoyed great success for boosting LLM performance on reasoning tasks without the need for any re-training. However, existing methods follow a think-first-then-decide paradigm, i.e., they first sample multiple reasoning paths, which inevitably leads to overgeneration, then prune or stop unnecessary paths to compensate. In contrast, decide-first-then-think, i.e., first identifying points that are likely to lead to...

    arxiv.org/abs/2606.31484 · PDF

  35. 35

    Constrained Online Convex Optimization without Slater's Condition

    Kihyun Yu, Junehee Lee, Dabeen Lee

    cs.LG · math.OC

    We study constrained online convex optimization with adversarial losses and stochastic or adversarial constraints. For stochastic constraints, existing algorithms that achieve nearly optimal regret and constraint violation bounds typically rely on regularity assumptions such as Slater's condition, while adversarial-constraint algorithms avoid these assumptions by using a rather restrictive round-wise feasible comparator. We bridge this gap...

    arxiv.org/abs/2606.31480 · PDF

  36. 36

    TabPATE: Differentially Private Tabular In-Context Learning Without Public Data

    Dariush Wahdany, Matthew Jagielski, Jesse C. Cresswell, Adam Dziedzic, Franziska Boenisch

    cs.LG

    Tabular foundation models enable accurate in-context learning (ICL) from small labeled datasets, but the private records placed in context can leak through model predictions. We first show that even basic membership inference attacks succeed against tabular ICL, motivating formal privacy protection. We then introduce TabPATE, a differentially private PATE-style defense for tabular ICL that does not require public in-distribution data. TabPATE...

    arxiv.org/abs/2606.31474 · PDF

  37. 37

    Zero-Shot Quantization for Object Detectors using Off-the-Shelf Generative Models

    Hyunho Lee, Kyomin Hwang, Hyeonjin Kim, Suyoung Kim, Sunghyun Wee, Nojun Kwak

    cs.LG

    With an increasing number of Object Detection (OD) models being deployed on edge devices, Zero-Shot Quantization for OD (ZSQ-OD) aims to quantize these models when access to the original training data is prohibited. Existing research on Zero-Shot Quantization-Aware Training (QAT) for OD synthesizes training sets through noise optimization. However, this approach struggles to maintain performance in low-bit regions. In this paper, we introduce...

    arxiv.org/abs/2606.31456 · PDF

  38. 38

    Contextual Slate GLM Bandits with Limited Adaptivity

    Tanmay Goyal, Sukruta Prakash Midigeshi, Gaurav Sinha

    cs.LG · stat.ML

    We investigate the contextual slate bandit problem with generalized linear rewards under limited adaptivity. At each round, the learner is presented with $N$ sets of items, where each item is represented by a $d$-dimensional feature vector. The learner then constructs a slate by selecting one item per set; the resulting slate yields a scalar reward sampled from a Generalized Linear Model (GLM). We propose algorithms under two...

    arxiv.org/abs/2606.31449 · PDF

  39. 39

    Mixture-of-Control: State-Aware Fine-Tuning for Transformer-based Models

    Duc Anh Nguyen, Tien Ngoc Luu, Tung Pham, Toan Tran

    cs.LG · cs.AI

    State-based fine-tuning has emerged as a compelling alternative to weight-based adaptation for transformers, updating lightweight controls into states rather than model weights, offering substantial memory savings while retaining parameter efficiency. However, most existing state-based methods typically apply only per-block control updates, which limits inter-block information exchange and restricts representational adaptation. Meanwhile,...

    arxiv.org/abs/2606.31397 · PDF

  40. 40

    Resolving superposition in AI for interpretability and cross-modal alignment in patient-neuronal images

    Jisung Park, Seohyeon Kang, Daeun Yoo, Eunsu Lee, Seoin Cho, Wooyeop Choi, Ian Choi, James R. Evan, Daesoo Kim,...

    cs.LG · cs.AI · cs.CV · q-bio.QM

    Artificial intelligence is transforming our capability to solve biological challenges. In dimensionality bottleneck regimes exacerbated by high-dimensional biological data, Neural networks force distinct concepts into the lower dimensions known as superposition. Although this superposition is widely known to hinder interpretability, its impact on corrupting the geometry of latent spaces remains critically overlooked. Here, we utilized sparse...

    arxiv.org/abs/2606.31394 · PDF

  41. 41

    Calibrating the Evaluator: Does Probability Calibration Mitigate Preference Coupling in LLM Agent Feedback Loops?

    Zewen Liu

    cs.LG · cs.AI · cs.CL

    When large language model (LLM) agents adapt their behavior through evaluator feedback, systematic evaluator biases propagate into the agent's learned strategy distribution - a phenomenon termed evaluator preference coupling. Prior work has documented this coupling and established a diagnostic framework (EPC) to measure it, but has not investigated whether calibration techniques can mitigate the effect. We present the first study of evaluator...

    arxiv.org/abs/2606.31371 · PDF

  42. 42

    Dualformer: Efficient Feature Extractor for Complex-valued Blind Communication Signal Analysis

    Yurui Zhao, Xiang Wang, Jingreng Lei, Wanlong Zhang, Yik-Chung Wu, Zhitao Huang

    cs.LG · eess.SP

    Designing effective feature extractors is critical for blind signal analysis tasks such as automatic modulation recognition (AMR), signal scheme recognition (SSR), and \color{black} signal structure parsing (SSP). In this work, we propose dual-channel neural network (DualNN) that efficiently exploits complex-valued signals through parameter sharing across IQ channels. Unlike traditional real-valued or complex-valued models, DualNN is a...

    arxiv.org/abs/2606.31352 · PDF

  43. 43

    Expected Gain-based Escalation in Vertical Federated Learning

    Mohamad Mestoukirdi, Vincent Corlay

    cs.LG

    Collaborative inference can improve predictive performance by integrating complementary information across agents, but applying collaborative fusion to every sample can incur unnecessary communication and computational overhead. This trade-off is particularly relevant in vertical federated learning (VFL), where clients observe different views of the same sample and fusion typically requires transmitting intermediate representations to a...

    arxiv.org/abs/2606.31331 · PDF

  44. 44

    Safe Online Learning via Smooth Safety-Structured Policy Composition

    Hongpeng Cao, Liqun Zhao, Yuliang Gu, Naira Hovakimyan, Lui Sha, Marco Caccamo

    cs.LG · cs.RO

    Safe online reinforcement learning requires policies to respect safety constraints while maintaining smooth optimization dynamics. Existing approaches typically rely on either strict safety enforcement via action interventions, which introduce discontinuities in system interaction and learning, or soft safety constraint formulations, which preserve smooth learning but provide limited safety assurance. We propose AutoSafe, a safety-aware...

    arxiv.org/abs/2606.31320 · PDF

  45. 45

    Deep Reinforcement Learning for Spacecraft Attitude Control During Atmospheric Re-Entry

    Alexander Fabisch, Melvin Laux, Mariela De Lucas Álvarez, Edoardo Caroselli, Julian Theis

    cs.LG

    Deep reinforcement learning has the potential to solve attitude control problems more adaptively, precisely, and robustly by handling nonlinear dynamics, uncertainties, and failure cases more effectively than traditional attitude control approaches. We explore reinforcement learning (RL) for attitude control in spacecraft re-entry. An industry-standard proportional-integral-derivative controller with gain scheduling serves as a strong...

    arxiv.org/abs/2606.31291 · PDF

  46. 46

    Patch-PODiff-ViT: Structured Latent Diffusion with Patchwise POD for Super-Resolution and Uncertainty Quantification

    Onkar Jadhav, Tim French, Matthew Rayson, Nicole L. Jones

    cs.LG

    Diffusion models enable probabilistic super-resolution and conditional generation, but pixel-space methods are computationally expensive and learned latent spaces often lack interpretable uncertainty quantification. We introduce Patch-PODiff-ViT, a structured latent diffusion framework in which the latent space is defined by patchwise Proper Orthogonal Decomposition (POD), a fixed linear orthonormal basis over local patches, rather than...

    arxiv.org/abs/2606.31290 · PDF

  47. 47

    Probabilistic Inversion with Flow Matching

    Baldur Paulwitz, Stefan Buske

    cs.LG · math.PR · physics.geo-ph

    We demonstrate the application of Flow Matching, a technique originating from generative Artificial Intelligence, to probabilistic inversion in geophysical settings, such as seismic Full-Waveform inversion. We adapt the well-established mathematical theory of Flow Matching from generative Artificial Intelligence to the context of probabilistic inversion. We evaluate the approach with two case studies: a simple 2D velocity model to illustrate...

    arxiv.org/abs/2606.31288 · PDF

  48. 48

    Sequential sparse Gaussian process quantile regression

    Hugo Nicolas, Olivier Le Maître

    cs.LG · stat.ML

    Quantile regression aims to estimate the conditional quantiles of a response variable from observed data. In a Bayesian setting, Gaussian process quantile regression provides uncertainty quantification but faces significant computational challenges due to the nonconjugacy of the asymmetric Laplace likelihood and the cost of posterior inference. We develop a sparse Gaussian process framework in which the quantile function is represented...

    arxiv.org/abs/2606.31284 · PDF

  49. 49

    Revisiting the Volume Hypothesis

    Ari Pakman, Lior Kreimer, Yakir Berchenko

    cs.LG

    Modern deep neural networks often contain far more parameters than needed to fit their training data, yet they achieve impressive generalization. A common explanation for this success is the implicit bias of stochastic gradient descent (SGD). An alternative volume hypothesis posits that, within low training-loss regions, loss-landscape basins leading to strong generalization occupy much larger regions of weight space than basins that...

    arxiv.org/abs/2606.31282 · PDF

  50. 50

    The Calibration Turn in AI-Assisted Research: A Conceptual and Methodological Framework for Evidence-Licensed Claims

    Hongmin Li

    cs.LG

    AI-assisted research has entered a stage in which the central question is not only whether systems can generate hypotheses, run experiments, or produce manuscripts, but whether their scientific claims are calibrated to the evidence that supports them. This Perspective-style paper develops a conceptual and methodological framework for evidence-licensed claims in AI-assisted research. Motivated by representative routes including specialized...

    arxiv.org/abs/2606.31273 · PDF

  51. 51

    TDGT: A Tabular Data Generation Toolkit supporting adaptive GPU-accelerated Bayesian mixture models, diffusion-based models, and latent-space generative modeling

    Vasileios C. Pezoulas, Nikolaos S. Tachos, Eleni Georga, Kostas Marias, Manolis Tsiknakis, Dimitrios I. Fotiadis

    cs.LG · cs.AI

    The growing demand for privacy-preserving data sharing has positioned synthetic data generation as a critical component of responsible AI workflows. Despite notable advances in generative modeling, existing solutions often lack integration of adaptive generation strategies, multi-metric evaluation, and accessible end-to-end generators within a unified web-based toolkit. In this work, we introduce TDGT (Tabular Data Generation Toolkit), a...

    arxiv.org/abs/2606.31268 · PDF

  52. 52

    Learning Gaussian Graphical Models from a Glauber Trajectory Without Mixing

    Eric Shen, Tony Wu, Mahbod Majid, Ankur Moitra

    cs.LG · stat.ML

    We study the task of learning the structure of a $d$-sparse Gaussian graphical model on $n$ variables from a single trajectory of Glauber dynamics. Beyond algorithmic considerations, many applications present temporally correlated observations rather than i.i.d.\ samples. In the classical i.i.d.\ setting, under comparably general sparsity and minimum edge-strength assumptions, sublinear-in-$n$ sample guarantees are known, but achieving them...

    arxiv.org/abs/2606.31230 · PDF

  53. 53

    Probing Memorization of Tabular In-Context Learning

    Francesco Capano, Jonas Böhler

    cs.LG · cs.CR

    Large tabular models (LTMs), i.e., tabular foundation models leveraging in-context learning (ICL), achieve state-of-the-art performance on tabular tasks. While LLMs are known to unintentionally memorize training data, the memorization dynamics of LTMs remain largely unexplored. We investigate the potential for parametric memorization in tabular ICL. We introduce ICLMEM, a probing framework designed to separate context-based predictions from...

    arxiv.org/abs/2606.31208 · PDF

  54. 54

    ISM:Self-Improving Strategy Memory for Continual Mathematical Reasoning

    Prakhar Dixit, Tim Oates

    cs.LG

    We propose Intelligent Schema Memory (ISM), a self-evolving memory-augmented system that improves mathematical reasoning for a frozen LLM under continual learning with hard episodic resets. ISM maintains a compact, self-refined bank of strategy schemas learned from both successful and failed episodes, with symbolic tools that check intermediate steps and certify answers.Without updating model parameters, ISM outperforms passive, retrieval,...

    arxiv.org/abs/2606.31191 · PDF

  55. 55

    Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation

    Jiachun Li, David Simchi-Levi

    cs.LG · cs.AI

    Adaptive experiments for average treatment effects (ATE) require randomized allocations balancing valid inference with statistical efficiency. The oracle design is a covariate-dependent Neyman rule governed by unknown arm-conditional outcome variances. We investigate whether this sequential variance-estimation and allocation process can be amortized via in-context learning. We introduce Bayesian in-context experimenters: transformer policies...

    arxiv.org/abs/2606.31184 · 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.