cs.LG · 2026-06-17 · No. 26

Machine Learning, 2026-06-17.

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

    Sign-Rank, Index, and List Replicability: Connections and Separations

    Ari Blondal, Hamed Hatami, Pooya Hatami, Chavdar Lalov, Sivan Tretiak

    cs.LG · cs.IT

    In learning theory, the sign rank of a binary concept class captures the smallest dimension in which it can be represented by points and halfspaces. Despite tremendous interest, lower bounds on sign rank are notoriously difficult to come by. Two recent approaches to the problem establish lower bounds on sign rank by measures that are easier to analyze: the $\mathbb{Z}_2$-index and the list replicability number. We order these measures,...

    arxiv.org/abs/2606.18236 · PDF

  2. 02

    Rethinking Dataset Distillation for Classification: Do Distilled Sets Outperform Coresets?

    Trisha Mittal, Akshay Mehra, Joshua Kimball

    cs.LG

    Dataset distillation (DD) has emerged as a prominent approach in data centric machine learning, aiming to synthesize compact training sets for efficient training by compressing the information in large datasets into a small number of synthetic samples. However, DD methods are often evaluated under inconsistent evaluation protocols, ranging from standard ERM to single/multi-teacher supervision, making it difficult to isolate the effectiveness...

    arxiv.org/abs/2606.18209 · PDF

  3. 03

    Looped World Models

    Hongyuan Adam Lu, Z. L. Victor Wei, Qun Zhang, Jinrui Zeng, Bowen Cao, Lingwei Meng, Mocheng Li, Zezhong Wang,...

    cs.LG · cs.AI · cs.CL · cs.CV

    Current world models face a fundamental tension: faithful long-horizon simulation demands deep computation, but deeper models are expensive to deploy and prone to compounding errors. We resolve this by introducing Looped World Models (LoopWM), which are the first looped architectures for world modelling. Our method iteratively refines latent environment states through a parameter-shared transformer block. This yield up to 100x parameter...

    arxiv.org/abs/2606.18208 · PDF

  4. 04

    Kolmogorov Regression for Robust Diffusion Policies

    Lekan Molu

    cs.LG · cs.AI

    Finite-dimensional (FD) diffusion policies exhibit temporal drift owing to discretization artifacts that degrade long-horizon performance (when deployed on physical systems). We introduce a backward Kolmogorov equation that lifts diffusion policies to a Cameron-Martin space -- a subset of the Hilbert space. Essentially, replacing stochastic score matching with a deterministic boundary-value PDE problem. Our core innovation thrives on Gaussian...

    arxiv.org/abs/2606.18186 · PDF

  5. 05

    Embedded Machine Learning for Microcontroller-Class Edge Devices: Data, Feature, Evaluation, and Deployment Pipelines

    Mostafa Darvishi

    cs.LG · cs.AI · cs.AR · eess.AS · eess.SP

    Embedded machine learning moves inference from cloud services to resource-constrained devices that must acquire data, preprocess signals, run a model, and act within tight limits on memory, energy, and latency. This paper presents a systems-oriented synthesis of an embedded machine-learning workflow for microcontroller-class platforms. The emphasis is placed on engineering decisions that are often hidden in generic machine-learning...

    arxiv.org/abs/2606.18122 · PDF

  6. 06

    Ternary Mamba: Grouped Quantization-Aware Training of W1.58A16 State Space Models

    Ramprasath Ganesaraja, Sahil Dilip Panse, Swathika N

    cs.LG · cs.AI

    State Space Models (SSMs) such as Mamba-2 offer linear-time inference but their memory footprint limits edge deployment. Prior ternary SSM work (Slender-Mamba) trains from scratch on 150B tokens; we show a pretrained checkpoint suffices, reducing the marginal token budget by 1,000x. Using grouped quantization-aware training (QAT) with knowledge distillation from a frozen FP16 teacher, we compress Mamba-2 1.3B to 3.61x (2,687 to 744 MB) and...

    arxiv.org/abs/2606.18114 · PDF

  7. 07

    Learning Fair Pareto-Optimal Policies in Multi-Objective Reinforcement Learning

    Umer Siddique, Peilang Li, Yongcan Cao

    cs.LG · cs.AI

    Fairness is an important aspect of decision-making in multi-objective reinforcement learning (MORL), where policies must ensure both optimality and equity across multiple, potentially conflicting objectives. While single-policy MORL methods can learn fair policies for fixed user preferences using welfare functions such as the generalized Gini welfare function (GGF), they fail to provide the diverse set of policies necessary for dynamic or...

    arxiv.org/abs/2606.18111 · PDF

  8. 08

    Deep Reinforcement Learning for Minimum Zero-Forcing Sets

    Steve Halley, Maurício Gruppi

    cs.LG

    This paper explores the problem of finding the minimum zero-forcing set on undirected graphs and proposes an adapted machine-learning framework to solve the problem. The minimum zero-forcing set problem is a graph coloring problem where the color of an initial set of nodes propagates throughout a network. The set of nodes is zero-forcing if it forces all uncolored nodes to change color under the constraint of the color-change rule. There are...

    arxiv.org/abs/2606.18106 · PDF

  9. 09

    S4oP: Operator-level Pruning of Structured State Space Models for Resource-Constrained Devices

    Marco Deano, Filippo Ziche, Nicola Bombieri

    cs.LG · cs.AI · cs.DC

    Structured State Space Models (SSMs), including the S4 and S4D architectures, have recently emerged as powerful alternatives to attention-based models for capturing long-range dependencies in sequential data. Despite their strong empirical performance, deploying these models in time- and resource-constrained settings remains challenging due to their computational and memory demands. In this paper, we propose a novel incremental,...

    arxiv.org/abs/2606.18096 · PDF

  10. 10

    From Reasoning Traces to Reusable Modules: Understanding Compositional Generalization in Language Model Reasoning

    Lingjing Kong, Xin Liu, Guangyi Chen, Martin Q. Ma, Xiangchen Song, Yuekai Sun, Mikhail Yurochkin, Taylor W....

    cs.LG

    Post-training pipelines that combine supervised fine-tuning (SFT) with reinforcement learning (RL) have emerged as the key recipe for transforming large language models (LLMs) into robust reasoners. We argue that this combined success is driven by compositional generalization, which we formalize through a hierarchical latent selection model. In this framework, reasoning traces are generated by a cascade of discrete latent selection variables...

    arxiv.org/abs/2606.18089 · PDF

  11. 11

    Edge Flow: A Tractable and Predictive Continuous-Time Model for Gradient Descent at the Edge of Stability

    Pierre Marion

    cs.LG

    Gradient descent in deep learning may operate at the edge of stability (EoS), a regime in which the largest eigenvalue of the loss Hessian hovers near the stability threshold $2/η$, where $η$ is the learning rate. Classical analysis tools such as gradient flow and the descent lemma do not apply here, motivating the search for a continuous-time model valid at EoS. We propose Edge Flow, a system of three coupled ordinary differential equations...

    arxiv.org/abs/2606.18080 · PDF

  12. 12

    Volterra Generative Models

    Yusen Jia, Bingyan Han

    cs.LG · cs.AI

    Score-based diffusion models typically use Brownian perturbations, which provide tractable reverse-time dynamics but impose memoryless noising. We introduce Volterra generative models, a continuous-time score-based framework whose forward process injects path-dependent noise through fractional kernels. To handle the non-Markovian and non-semimartingale dynamics, we construct finite-dimensional Markovian lifts using Gaussian quadrature in both...

    arxiv.org/abs/2606.18071 · PDF

  13. 13

    NoiseTilt: Noise-Tilted Reverse Kernels for Diffusion Reward Alignment

    Jisung Hwang, Yunhong Min, Jaihoon Kim, I-Chao Shen, Minhyuk Sung

    cs.LG

    We introduce the Noise-Tilted Reverse Kernel (NTRK), a reward-guided diffusion sampler that injects reward gradients through the noise term, leaving the pretrained reverse kernel unchanged and requiring only a single sample per step. Reward-guided sampling at inference time has greatly expanded the versatility of pretrained diffusion models. Yet existing methods face a trade-off. Gradient-based guidance shifts the reverse mean, steering...

    arxiv.org/abs/2606.18066 · PDF

  14. 14

    ConTex: Reformulating Counterfactual Generation For Time Series Forecasting

    Jan Voets, Hasan Tercan, Tobias Meisen, Sebastian Baum

    cs.LG

    Decision-making with deep learning-based time series forecasting requires not only accurate predictions but also actionable insights. However, current architectures do not inherently provide such information. Specifically, guidance is needed on how current conditions must be modified to shift from a predicted outcome to a desired future scenario. Counterfactual explanations provide a natural framework for this task, as they represent minimal...

    arxiv.org/abs/2606.18049 · PDF

  15. 15

    Catastrophic Forgetting is Low-Rank: A Function-Space Theory for Continual Adaptation

    Ido Nitzan Hidekel, Dan Raviv

    cs.LG · cs.AI

    Catastrophic forgetting in continual adaptation is usually studied through parameter drift, replay, or distillation, but these views do not identify which output-space directions are vulnerable. We give a function-space account in the NTK regime: new-task training induces old-task prediction drift through the cross-task kernel, yielding a closed-form predictor for the forgetting vector before any new-task gradient step. In frozen-backbone...

    arxiv.org/abs/2606.18024 · PDF

  16. 16

    LoopCoder-v2: Only Loop Once for Efficient Test-Time Computation Scaling

    Jian Yang, Shawn Guo, Wei Zhang, Tianyu Zheng, Yaxin Du, Haau-Sing Li, Jiajun Wu, Yue Song, Yan Xing, Qingsong Cai,...

    cs.LG · cs.AI

    Looped Transformers scale latent computation by repeatedly applying shared blocks, but sequential looping increases latency and KV-cache memory with the loop count. Parallel loop Transformers (PLT) alleviate this cost through cross-loop position offsets (CLP) and shared-KV gated sliding-window attention, making loop count a practical design choice. We therefore study PLT loop-count selection through a gain--cost view: an extra loop may refine...

    arxiv.org/abs/2606.18023 · PDF

  17. 17

    Recursive Scaling in Masked Diffusion Models

    Alba Carballo-Castro, Julianna Piskorz, Paulius Rauba, Mihaela van der Schaar, Pascal Frossard

    cs.LG

    Masked diffusion models (MDMs) have recently emerged as a promising paradigm for sequence generation. Scaling MDMs is conventionally achieved by increasing the parameter count or the number of denoising steps. We introduce Recursive Masked Diffusion Models (R-MDMs), which add recursive depth as a third scaling axis by repeatedly applying the same denoising transformer within each diffusion step. Recursion enables iterative refinement of the...

    arxiv.org/abs/2606.18022 · PDF

  18. 18

    C2FL: Clustered Continual Federated Learning under Spatial and Temporal Drift

    Davide Domini, Gianluca Aguzzi, Lorenzo Pellegrini, Mirko Viroli, Lukas Esterle

    cs.LG · cs.AI

    Collective Adaptive Systems (CAS) increasingly rely on machine learning to let each node learn from locally sensed data, aligning its behavior with the surrounding environment. Scaling this intelligence, however, raises fundamental challenges: sensed data is often privacy-sensitive, preventing centralized collection; nodes are mobile, traversing regions where nearby nodes perceive similar phenomena while distant ones observe radically...

    arxiv.org/abs/2606.18003 · PDF

  19. 19

    Half a Link can Be Enough to Predict a Whole Link: Understanding Generalization in Knowledge Graph Foundation Models

    Cosimo Gregucci, Obaidah Theeb, Daniel Hernandez, Antonio Vergari, Steffen Staab

    cs.LG

    Knowledge graph (KG) foundation models (KGFMs) are zero-shot generalizers: trained once, they can predict links on unseen graphs without retraining. However, understanding when and how they can robustly generalize across KGs is still an open question. In this paper, we shed some light on their generalization mechanisms highlighting how their performance on unseen KGs is not uniform when it comes to partially seen links, which we call...

    arxiv.org/abs/2606.18001 · PDF

  20. 20

    Multiple cyclicity and Wavelet Decomposition with Channel Correlation for Long-term Time Series Forecasting

    Bin Wang, Heming Yang, Jinfang Sheng

    cs.LG · cs.AI

    Cyclicity and trend are important components of time series data and many studies based on cyclicity and trend have achieved good results in long-term time series forecasting. However, we believe that current work neglects the influence of real-world inter-channel correlations in time series data which leads to suboptimal predictions. Furthermore, these models rely on complex designs to capture diverse information so that resulting in low...

    arxiv.org/abs/2606.17996 · PDF

  21. 21

    SoftMoE: Soft Differentiable Routing for Mixture-of-Experts in LLMs

    Mikołaj Zasada, Łukasz Struski, Jacek Tabor, Marcin Kurdziel

    cs.LG · cs.AI

    Sparse Mixture-of-Experts (MoE) architectures enable scaling LLM parameters under a fixed inference budget by activating only a small subset of experts via top-$k$ routing. While this preserves causality and suits autoregressive language models, the discrete top-$k$ operator is not differentiable, forcing a fixed number of active experts per input and resulting in inefficient use of computation. We propose SoftMoE, which replaces discrete...

    arxiv.org/abs/2606.17952 · PDF

  22. 22

    Predictive Analytics in E-Commerce for CustomerBehavior Forecasting using hybrid Ret-DNN withXGBoost Model

    Degala Pushpa Sri, Mayank Atreya, Lakshmi. H, Navin Chhibber, Mukesh Soni

    cs.LG

    In recent years, electronic (E) commerce services have rapidly increased in the daily lives of people, which helpsthem to purchase products online. However, retail platforms have struggled to understand customer behavior and make it difficult to predict their future purchases. To overcome these challenges, this study proposes a hybrid Retail Deep NeuralNetwork (Ret-DNN) with an Extreme Gradient Boosting(XGBoost) model for capturing temporal...

    arxiv.org/abs/2606.17931 · PDF

  23. 23

    KANLib -- An Modular, Extensible and Fast Kolmogorov-Arnold Network Implementation

    Julian Hoever, Gregor Schiele

    cs.LG · cs.AI

    Kolmogorov-Arnold Networks (KANs) have recently emerged as a promising alternative to traditional multilayer perceptrons by replacing linear weights with learnable univariate functions. Despite their theoretical advantages in interpretability and expressiveness, practical research of KANs remains difficult due to high computational costs and inconsistent feature support across existing frameworks. This paper introduces KANLib, a modular,...

    arxiv.org/abs/2606.17927 · PDF

  24. 24

    Dimensionality Controls When Modularity Helps in Continual Learning

    Kathrin Korte, Christian Medeiros Adriano, Joachim Winther Pedersen, Eleni Nisioti, Sebastian Risi

    cs.LG · cs.AI · cs.NE

    Compositional learning systems must balance plasticity, the ability to acquire new knowledge, with stability, the preservation of previously learned components, especially when tasks share structure and risk interference. We study how modular architecture, task similarity, and representational dimensionality jointly shape compositional continual learning in a sequential A-B-A paradigm, comparing a task-partitioned recurrent network to a...

    arxiv.org/abs/2606.17889 · PDF

  25. 25

    Monotonic Kolmogorov-Arnold Networks: A Theoretical and Empirical Study of Monotonicity as an Inductive Bias

    Mikhail Krasnov, Carolina Fortuna, Blaž Bertalanič

    cs.LG

    Monotonicity has been a long-running architectural inductive bias for neural networks, motivated by tabular, scientific, and economic settings where outputs are known to respond monotonically to certain inputs. Existing approaches are MLP- or flow-based and lack per-edge functional transparency; the only Kolmogorov--Arnold Network (KAN) variant with monotonicity, MonoKAN, enforces the constraint only on a restricted parameter subset and...

    arxiv.org/abs/2606.17886 · PDF

  26. 26

    AnchorKV: Safety-Aware KV Cache Compression via Soft Penalty with a Refusal Anchor

    Ning Ni, Yingjie Lao

    cs.LG · cs.AI

    Large language models (LLMs) outperform earlier architectures on generative inference and long-context tasks, but their large size introduces significant challenges in memory usage, energy cost, and on-device deployment. Since scaling pre-trained language models improves downstream capability \cite{zhao2023survey}, the key-value (KV) cache becomes a dominant inference bottleneck. Recent KV cache compression methods...

    arxiv.org/abs/2606.17872 · PDF

  27. 27

    Meta-classification of one-class classification models using ranking correlation and nearest neighbor

    Toshitaka Hayashi, Hamido Fujita, Dalibor Cimr, Richard Cimler, Jitka Kühnová

    cs.LG

    Machine Learning (ML) techniques have been applied to various problems. However, applying ML to ML models is an unexplored direction. For this purpose, this paper considers a meta-classification of one-class classification (OCC) models, because all ML models could be approximated as OCC models. The proposal represents OCC models as normality rankings and classifies them using nearest-neighbor and ranking-correlation metrics. The experiment...

    arxiv.org/abs/2606.17858 · PDF

  28. 28

    From Drift to Coherence: Stabilizing Beliefs in LLMs

    SongEun Kim, Seungyoo Lee, Edwin Fong, Hyungi Lee, Juho Lee

    cs.LG

    Large language models (LLMs) are often hypothesized to perform implicit Bayesian inference, yet a key coherence condition, the martingale property of predictive beliefs, has been shown to fail in controlled synthetic in-context learning settings. We revisit this question in a more typical usage regime: generic multiple-choice question answering. Exploiting the discrete answer space, we compute exact predictive distributions and study belief...

    arxiv.org/abs/2606.17832 · PDF

  29. 29

    Functional Equivalence in Attention: A Comprehensive Study with Applications to Linear Mode Connectivity

    Viet-Hoang Tran, Vinh Khanh Bui, Van-Hoan Trinh, Tan Lai Ngoc, Tan M. Nguyen

    cs.LG · cs.AI

    Neural network parameter spaces are inherently non-injective, as distinct parameter configurations can realize identical functions through functional equivalence. While this symmetry is well understood in classical fully connected and convolutional models, it becomes substantially more intricate in modern attention-based architectures. Existing analyses of multihead attention have largely focused on the vanilla formulation, overlooking...

    arxiv.org/abs/2606.17830 · PDF

  30. 30

    Conservation Laws for Modern Neural Architectures

    Viet-Hoang Tran, Vinh Khanh Bui, Tan Lai Ngoc, Nam Nguyen, Tuan Dam, Tan M. Nguyen

    cs.LG · cs.AI

    Understanding gradient descent dynamics is key to explaining the success of over-parameterized models, where implicit bias manifests through conservation laws in gradient flow. While such laws are well understood for linear and ReLU networks, they remain largely unexplored for modern architectures. This work develops a unified framework to characterize conservation laws for contemporary models, including feedforward networks with GELU, SiLU,...

    arxiv.org/abs/2606.17816 · PDF

  31. 31

    No-Free-Fairness: Fundamental Limits and Trade-offs in Learning Systems

    Khoat Than

    cs.LG · cs.AI

    In this paper, we establish a set of theoretical impossibility results, termed the No-Free-Fairness theorems, that identify three fundamental sources of disparity in learning systems. First, we show that when a task exhibits irreducible cost on a subgroup, any decision rule must trade off overall performance with disparity, yielding an inherent fairness--cost frontier. Second, we prove that even in ideal, noise-free settings where a perfectly...

    arxiv.org/abs/2606.17810 · PDF

  32. 32

    QueryMarket: Cost-Aware Online Active Learning in Data Markets

    Xiwen Huang, Pierre Pinson

    cs.LG

    Data acquisition is a major bottleneck for learning in real-time streams: analysts must decide on the fly which labels to purchase while respecting a rolling budget. However, existing online active learning rarely unifies pricing, information gain, and rolling budget constraints under concept drift. We introduce QueryMarket, a market-inspired framework that queries each incoming data point based on its estimated utility to the model and its...

    arxiv.org/abs/2606.17805 · PDF

  33. 33

    Continual Self-Improvement with Lightweight Experiential Latent Memories

    Vaggelis Dorovatas, Nancy Kalaj, Rahaf Aljundi

    cs.LG

    Large language models achieve strong reasoning performance by scaling inference-time compute, yet remain fundamentally stateless, discarding the rich, self-produced reasoning traces generated during this process. We investigate whether models can instead learn online from this experience, converting transient computation (reasoning traces) into persistent reusable knowledge, and without external supervision or access to future data. We show...

    arxiv.org/abs/2606.17803 · PDF

  34. 34

    Blind Recovery of Latent Domains via Unsupervised Symmetry Discovery

    Onur Efe, Arkadas Ozakin

    cs.LG

    Primary motivation in blind inverse problems is to recover signals of interest from corrupted observations without knowing the obfuscating mechanism. Blind deconvolution is a prominent approach when the corruption is convolutional, but it is not applicable when general linear transformations obfuscate the domain structure. In this work, we propose an unsupervised framework for recovering latent domains and signals by discovering symmetries of...

    arxiv.org/abs/2606.17782 · PDF

  35. 35

    A fairness-aware extension of Stochastic Multicriteria Acceptability Analysis for ranking

    Guilherme Dean Pelegrina, Renata Pelissari

    cs.LG

    Fairness has become a central concern in ranking problems involving individuals or social groups, particularly under the Responsible Artificial Intelligence agenda. In Multi-Criteria Decision Analysis, Stochastic Multicriteria Acceptability Analysis (SMAA) provides a robust framework for handling uncertainty and incomplete preference information, but it does not explicitly address fairness in the resulting rankings. This paper proposes...

    arxiv.org/abs/2606.17756 · PDF

  36. 36

    Confusion-Aware Transfer Teacher Curriculum Learning Framework: Disentangling Scoring and Pacing Effects

    Savini Kommalage, Sanka Mohottala, Asiri Gawesha, Dulara Madhusanka, Menan Velayuthan, Dharshana Kasthurirathna,...

    cs.LG · cs.AI

    Curriculum learning couples two design choices, how samples are scored by difficulty and how harder samples are paced into training, making it difficult to attribute observed gains to either component. We disentangle these factors with two evaluation protocols: stage-wise test subsets that validate scoring functions independently of curriculum training, and a baseline that applies the same pacing schedule to randomly ordered data. Within the...

    arxiv.org/abs/2606.17706 · PDF

  37. 37

    Delta-Based Target Reformulation for Short-Term Electricity Load Forecasting Using LSTM and Transformer Models

    Vansh Bansal

    cs.LG

    Accurate short-term electricity load forecasting is critical for the reliable and economic operation of modern power systems, under non-stationarity arising from weather variability, calendar effects, and evolving consumption patterns. While deep learning models such as LSTMs and Transformers show promising performance, most existing studies focus on direct absolute load prediction without explicitly addressing target non-stationarity....

    arxiv.org/abs/2606.17692 · PDF

  38. 38

    EnvRL: Learn from Environment Dynamics in Agentic Reinforcement Learning

    Zhitong Wang, Songze Li, Hao Peng, Shuzheng Si, Yi Wang, Maosong Sun, Juanzi Li

    cs.LG · cs.CL

    Reinforcement learning (RL) has emerged as a powerful paradigm for training Large Language Models (LLMs) as agents. However, conventional RL methods for long-horizon agentic tasks often struggle with sparse outcome rewards. Intuitively, this overlooks the rich environment dynamics information contained in rollout interaction trajectories. We argue that the interaction experience inherently serves as an implicit supervision signal, reveals the...

    arxiv.org/abs/2606.17680 · PDF

  39. 39

    ASTEROID: A Spatiotemporal Information Transformer for Forecasting Multi-Step Time Series of Molecular Dynamics

    Kexin Wu, Luonan Chen, Renxiao Wang

    cs.LG · cs.AI · q-bio.QM

    Molecular dynamics (MD) simulation is computationally demanding, particularly for large-scale systems requiring long-term analysis. Accurate forecast of the outcomes of a MD simulation is not only an attractive scientific challenge but also has substantial practical value. In this work, we developed a data-driven framework, termed ASTEROID (Advanced Spatiotemporal TransformER fOr Inferring Dynamics), that can directly predict multi-step...

    arxiv.org/abs/2606.17668 · PDF

  40. 40

    Handling Feature Heterogeneity with Learnable Graph Patches

    Yifei Sun, Yang Yang, Xiao Feng, Zijun Wang, Haoyang Zhong, Chunping Wang, Lei Chen

    cs.LG · cs.AI

    In recent years, the rapid development of foundation models and graph pre-training technologies has spurred increasing interest in constructing a universal pre-trained graph model or Graph Foundation Model (GFM). However, a significant challenge is that existing models are unable to address feature heterogeneity in graph data without textual information, which hinders the transferability of graph models across different datasets. To bridge...

    arxiv.org/abs/2606.17667 · PDF

  41. 41

    TuneAhead: Predicting Fine-tuning Performance Before Full Training Begins

    Yuxiang Luo, Haonan Long, Chen Wang, Qiqi Duan, Xiaotian Lin, Yanwei Xu, Yuyu Luo, Weikai Yang, Nan Tang

    cs.LG · cs.AI

    Fine-tuning large language models (LLMs) is compute-intensive and error-prone: model performance depends sensitively on data quality and hyperparameter choices, and naïve runs can even degrade model performance. This raises a practical question:can we predict fine-tuning performance before committing to a full training run? We present TUNEAHEAD, a lightweight framework for pre-hoc prediction of fine-tuning performance. TUNEAHEAD encodes each...

    arxiv.org/abs/2606.17660 · PDF

  42. 42

    Physics-Constrained Neural Networks for Improved Short-Term Weather Forecasting: A Case Study over the South Pacific

    Egor Bugaev, Fedor Buzaev, Dmitry Efremenko, Denis Derkach, Fedor Ratnikov

    cs.LG

    This study introduces enhancements to physics-constrained neural networks (PCNNs) that improve the accuracy and stability of hybrid short-term weather forecasting models. Building on the WeatherGFT architecture, three innovations are proposed. First, an upgraded numerical solver, combining a fifth-order weighted essentially non-oscillatory scheme (WENO-5), a beta-plane approximation, and subgrid-scale viscosity, permits a fourfold increase in...

    arxiv.org/abs/2606.17659 · PDF

  43. 43

    A Risk Decomposition Framework for Pre-Hoc Fine-Tuning Prediction

    Yuxiang Luo, Chen Wang, Nan Tang

    cs.LG · cs.AI

    The high cost of fine-tuning LLMs poses a significant economic barrier; pre-hoc performance prediction offers a critical solution to substantially reduce this expense. However, the theoretical limits of pre-hoc performance prediction remain unexplored. We formulate it as a stochastic estimation problem under information constraints, decomposing prediction risk into two components: an intrinsic limit (static data-model compatibility) and a...

    arxiv.org/abs/2606.17649 · PDF

  44. 44

    Expanding SPHERE-JEPA: A Family of Statistical Regularizers for the Hypersphere

    Léo Nicollier, Enric Meinhardt-Llopis, Max Dunitz, Marc Pic, Pablo Musé, Gabriele Facciolo

    cs.LG

    In Self-Supervised Learning (SSL), preventing representation collapse by explicitly enforcing a uniform distribution on the unit hypersphere has proven to be effective. However, current frameworks typically rely on sliced statistical regularizers such as SIGReg (used in LeJEPA) and SUSReg (used in SPHERE-JEPA), which approximate this continuous objective via Monte Carlo sampling along random 1D directions. This stochasticity injects...

    arxiv.org/abs/2606.17603 · PDF

  45. 45

    LLM Features Can Hurt GNNs: Concatenation Interference on Homophilous Graph Benchmarks

    Zhongyuan Wang, Pratyusha Vemuri

    cs.LG · cs.AI · cs.CL · cs.SI

    Adding LLM-generated node features to graph neural networks (GNNs) is widely reported to improve accuracy on standard benchmarks. We document a contrasting observation: when LLM features are introduced through pure input concatenation (rather than joint training, distillation, or prompt-conditioning), they can systematically degrade accuracy on the same homophilous benchmarks where end-to-end LLM pipelines succeed. With an MLP backbone on the...

    arxiv.org/abs/2606.17579 · PDF

  46. 46

    When Dynamics Models Read the Wrong Time Steps: Label-Free Event Credit Re-Anchoring for Robust Global Readouts

    Yifan Wang

    cs.LG · eess.SY

    Learned dynamics models often answer global physical questions, such as fault severity or impact stiffness, by pooling a per-step feature sequence into one readout vector. This sequence-to-global interface creates an under-studied temporal credit problem: with only trajectory-level supervision, a model can predict accurately in training conditions while reading from abundant smooth correlates rather than the brief physical events that...

    arxiv.org/abs/2606.17572 · PDF

  47. 47

    Reducing Learner Redundancy in Boosting via Residual Orthogonalization

    Ye Su, Jipeng Guo, Yong Liu, Xin Xu, Gangchun Zhang, Jinxin Chen, Di Wu, Longlong Zhao

    cs.LG

    While sequential residual fitting is the bedrock of standard boosting frameworks, it inherently breeds learner redundancy by repeatedly revisiting correlated error components. To address this bottleneck, we propose a shift from residual fitting to \textit{residual orthogonalization} and introduce SCBoost. Our framework tackles redundancy through two complementary mechanisms: Spectral Residual Projection (SRP) and Covariance-Regularized...

    arxiv.org/abs/2606.17567 · PDF

  48. 48

    SpatioTemporal Causal Network Diagnostics for Geographic Tipping Point Early Warning

    Zhaoyuan Yu, Zhangyong Liang

    cs.LG

    Geographic tipping points in ecosystems, climate subsystems, or ice sheets pose severe challenges for localized early warning. Classical spatial indicators such as Moran's I summarize global spatial structure, but they struggle with three issues: spatial dilution, Euclidean assumptions, and correlated noise. This paper introduces SpatioTemporal Causal Network Diagnostics (ST-CND), a framework that addresses these three issues by representing...

    arxiv.org/abs/2606.17553 · PDF

  49. 49

    Reversal Q-Learning

    Aditya Oberai, Seohong Park, Sergey Levine

    cs.LG · cs.AI

    Iterative generative modeling techniques, such as flow matching, provide powerful tools to model complex behaviors for effective offline reinforcement learning (RL). In this work, we propose a new off-policy RL algorithm that trains a flow policy based on prior data. Our idea starts from the "expanded" Markov decision process (MDP) framework, which treats individual flow refinement steps as separate actions in an MDP. To enable off-policy RL...

    arxiv.org/abs/2606.17551 · PDF

  50. 50

    Continuous-time Optimal Stopping through Deep Reinforcement Learning

    Cosmin Borsa, Michael Ludkovski

    cs.LG · q-fin.CP · q-fin.PR

    Simulation based solvers for optimal stopping problems must discretize the stopping decision. Under classical dynamic programming, a coarse exercise grid with only a few stopping opportunities can materially undervalue the optimal expected reward, whereas on a very fine grid, approximation errors accumulate through the backward recursion. To remove this limitation, we develop a new reinforcement-learning inspired algorithm that enables us to...

    arxiv.org/abs/2606.17545 · PDF

  51. 51

    Offline Preference-Based Trajectory Evaluation

    Fernando Diaz

    cs.LG · cs.AI

    Offline evaluation of agentic systems often collapses trajectories to terminal success, discarding information about partial progress and inducing widespread ties, creating substantial statistical inefficiency by reducing effective sample size and weakening the ability to distinguish systems. We propose preference-based trajectory evaluation, which compares trajectories directly through temporal preferences over progress and time-to-return...

    arxiv.org/abs/2606.17541 · PDF

  52. 52

    Non-negative Matrix Factorisation with Topological Regularisation

    Matias de Jong van Lier, Shizuo Kaji, Keunsu Kim

    cs.LG · cs.CG · math.AT

    We investigate the learning of interpretable bases in non-negative matrix factorisation (NMF) by regularising the topology of the learned basis functions. Our approach is motivated by the observation that many data modalities can be viewed as non-negative functions on a structured domain, where the quality of a basis is intrinsically linked to its topology. However, naive methods for incorporating the topology of the support are often...

    arxiv.org/abs/2606.17531 · PDF

  53. 53

    MGUP: A Momentum-Gradient Alignment Update Policy for Stochastic Optimization

    Da Chang, Ganzhao Yuan

    cs.LG

    Efficient optimization is essential for training large language models. Although intra-layer selective updates have been explored, a general mechanism that enables fine-grained control while ensuring convergence guarantees is still lacking. To bridge this gap, we propose \textbf{MGUP}, a novel mechanism for selective updates. \textbf{MGUP} augments standard momentum-based optimizers by applying larger step-sizes to a selected fixed proportion...

    arxiv.org/abs/2606.17526 · PDF

  54. 54

    Learning to Refine Hidden States for Reliable LLM Reasoning

    Chia-Hsuan Hsu, Jui-Ming Yao

    cs.LG

    Large language models show strong reasoning ability, but their internal reasoning process can remain unstable in complex multi-step settings, where early hidden-state errors may propagate to incorrect predictions. We propose ReLAR, a reinforcement-guided latent refinement framework that iteratively updates hidden representations before decoding. ReLAR maintains a compact latent reasoning state and uses learned depth and action controllers to...

    arxiv.org/abs/2606.17524 · PDF

  55. 55

    FoundCause: Causal Discovery with Latent Confounders from Observational Data

    Patrick Blöbaum, Krishnakumar Balasubramanian, Shiva Prasad Kasiviswanathan

    cs.LG · cs.AI · stat.ME · stat.ML

    Causal discovery from observational data remains challenging due to the need to recover directed structure and latent confounding without interventions. We propose FoundCause, an amortized causal discovery model trained entirely on synthetic data that maps datasets directly to causal graphs in a single forward pass. By learning from large collections of simulated structural causal models, FoundCause captures transferable statistical patterns...

    arxiv.org/abs/2606.17516 · PDF

  56. 56

    Geometry-Aware Post-Hoc Uncertainty Quantification in Operator Learning

    Oriol Vendrell-Gallart, Nima Negarandeh, Ramin Bostanabad

    cs.LG · cs.AI

    Neural operators provide fast surrogates for PDEs but their deterministic predictions limit their use in tasks requiring uncertainty quantification (UQ), especially under geometric variability. Existing approaches primarily model uncertainty in network parameters, largely overlooking the geometry-aware representations learned by the operator itself. We propose REEF-GP (Residual on Embedded Features Gaussian Process), a post-hoc UQ framework...

    arxiv.org/abs/2606.17513 · PDF

  57. 57

    When the Next Step Is Not One Step: Distribution-Aware Execution Modeling for Concurrent Go Programs

    Kaviru Hapuarachchi

    cs.LG · cs.DC · cs.PL · cs.SE

    Training a model to predict the next step in a concurrent program is harder than it looks: two runs of the same program from the same trace prefix can produce different next events, both valid, because the scheduler is nondeterministic. A model trained against a single label is learning to guess one outcome of a random process. We turn this around and use the nondeterminism as a training signal. We run each program many times, aggregate the...

    arxiv.org/abs/2606.17508 · PDF

  58. 58

    Reconfigurable Computing Challenge: Transformer for Jet Tagging on Versal AI Engines

    Gram Koski, Sean Lipps, Zhenghua Ma, G. Abarajithan, Ryan Kastner

    cs.LG · cs.AR

    Transformer-based models achieve strong performance for jet tagging at the CERN LHC, but deploying them in low-latency, resource-constrained trigger systems is challenging. We present an initial implementation of a quantized, integer-only transformer for jet tagging on the AMD Versal AI Engine (AIE), mapping dense and multi-head attention (MHA) layers to AIE tiles. The main contribution is a reusable software framework that represents...

    arxiv.org/abs/2606.17500 · PDF

  59. 59

    Online LLM Selection via Constrained Bandits with Time-Varying Demand

    Yin Huang, Qingsong Liu, Jie Xu

    cs.LG · cs.AI

    Large Language Models (LLMs) are increasingly deployed in edge-cloud inference systems to handle diverse user tasks with heterogeneous accuracy, latency, and cost profiles. Selecting the appropriate LLM for each incoming task is critical for ensuring service quality and efficient resource utilization. However, model heterogeneity, stochastic and unknown performance characteristics, and time-varying task demands make static selection...

    arxiv.org/abs/2606.17489 · 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.