AI News

Q-DEQ: Discrete Solving and Quantization for Deep Equilibrium Models in Time Series Forecasting under Edge Deployment Coding Constraints (arxiv.org)

FlashBoB: I/O-Efficient Exact Backward-over-Backward for Softmax Attention (arxiv.org)

CAMFT: Conflict-Aware Mergeable Fine-Tuning for Large Language Models (arxiv.org)

CALM: A Calibrated LLM Choice Network Framework for Activity-Based Traveler Simulation (arxiv.org)

Opinion Leader Dynamics: How Sparse Attention Shapes Token Clustering (arxiv.org)

H-Spec: Parallel Speculative Decoding Without a Drafter-Side KV Cache (arxiv.org)

Reinforcement Learning under State and Outcome Uncertainty: A Foundational Distributional Perspective (arxiv.org)

SPeaR: Test-Time Adaptation with Steering Primitives for Realigning Representations (arxiv.org)

PAC-Bayesian Meta-Learning for Few-Shot Identification of Linear Dynamical Systems (arxiv.org)

CLOOPD: Closing the Learner Loop in On-Policy Distillation (arxiv.org)

Luck Is Not Skill: When Do Paired Rollouts Help Group-Relative RL of LLM Agents? (arxiv.org)

Mind or Message? Auditing Theory of Mind in Multi-Agent Social Simulation (arxiv.org)

Acceptance-Aware Draft Model Training for Speculative Decoding (arxiv.org)

Discovering Physical Representation Languages (arxiv.org)

Blind Thermodynamic Ontology Discovery from Anonymous Experiments (arxiv.org)

Tool-Augmented On-Policy Distillation for LLM Domain Adaptation in Sequence-Based Omics Tasks (arxiv.org)

The Evidence Ladder for Reinforcement Learning in Healthcare: From Retrospective Policies to Trusted Interventions (arxiv.org)

What Can a Recurrent State Safely Forget? (arxiv.org)

FIRM-WM: State-factorized factual-interventional recurrent modeling for reward-free visual planning (arxiv.org)

Robust Market Making with Hawkes Order Flow and Price Impact via Adversarial Reinforcement Learning (arxiv.org)

Signal-Informed Temporal Routing for Vinyl Defect Regime Detection (arxiv.org)

SCALE: Simulation-Calibrated Amortized Learning for Energy Materials (A hybrid architecture connecting deterministic modeling, real-world data, and transformer-scale inference for accelerated energy-materials discovery) (arxiv.org)

List Counting Failures Are Not One Phenomenon (arxiv.org)

RLVR$^{2}$: Reinforcement Learning with Verifiable Rubric-based Ranking (arxiv.org)

Improved Private Sparse Covariance Estimation with Multiscale Threshold Tests (arxiv.org)

Look Before You Steer: Geometry Predicts SAE Feature Steerability (arxiv.org)

D-IMPL: A Diffusion-based Solver for Parameterized BBOs (arxiv.org)

Autonomous Model Lifecycle Management for Digital Twin-Based Manufacturing Control (arxiv.org)

TRACE: Tractable Routing Autoencoder for Clinical ECG (arxiv.org)

ITSY: Causal Discovery From Irregular Time-Series Data (arxiv.org)

A Synthetic Multivariate Refrigerator Time-Series Dataset for Predictive Maintenance (arxiv.org)

Can Coding Agents Reproduce Official Statistics? Metadata, Retry Budget and the Limits of Execution Feedback in a Controlled Eurostat Benchmark (arxiv.org)

Measuring the Checker: Mutation Analysis for GPU-Kernel Benchmark Oracles (arxiv.org)

Toollery: Scaling LLM Agents to Thousands of Skills and Tools (arxiv.org)

The Effect of Quantization on Clinical Benchmarks: Accuracy and Safety Across Model Families (arxiv.org)

Feature Suppression and Differential Privacy for Residential Traffic Classification: A Two-Home Federated Study (arxiv.org)

On Emergent Capabilities and Model Merging (arxiv.org)

Guaranteed Low-Rank Tensor Recovery from Modewise Measurements via Normalized Block-Weighted Riemannian Gradient Descent (arxiv.org)

Predicting Out-of-Distribution Generalization of Neural Operators via Observable Spectral Error Decomposition (arxiv.org)

Preserving Geometric Integrity in Graph Prompting via Measure-Constrained Optimal Transport (arxiv.org)

Dissecting Training-Free Uncertainty Estimation in Multimodal Large Language Models (arxiv.org)

Replicating the Geometry of Emotion Representations in a Base Open-Weights Model (arxiv.org)

Physics-residual machine learning predicts oxygen-evolution catalyst activity beyond the training range from sparse polarization measurements (arxiv.org)

Global Ranks Survive, Selected Heads Shift: BOS-Sink Topology under 4-bit Weight-Only Quantization (arxiv.org)

Cost-Aware Reinforcement Learning with Action Masking and Projection for Battery Energy Storage Dispatch under Suppressed-Spread Market Shifts (arxiv.org)

Bilinear Optimization Divergence: Diagnosing Factor-Constrained LoRA Continual Learning (arxiv.org)

ETH-TraceBench: A Large-Scale Event-Stream Benchmark for Ethereum DeFi under Temporal, Protocol, and Contract Shift (arxiv.org)

Predictors and Orchestrators: Parsimonious Machine Learning within an Agentic AI Harness for Multi-Horizon Karst Aquifer Forecasting (arxiv.org)

Hessian Rank Constraint for Learning Structure of Nonlinear Latent Variable Models (arxiv.org)

Do LLMs Choose Like Humans? Using Cognitive Theory to Evaluate LLM Decision-Making (arxiv.org)