AI News

Let It Go or Learn to Self-Correct: Continuous Diffusion for Constrained Discrete Tasks (arxiv.org)

Physics-Informed Deep Learning for False Ventricular Tachycardia Alarm Reduction in the ICU (arxiv.org)

Transformers as In-Context Samplers: From Closed-Form Diffusion to Estimation-Free Sampling (arxiv.org)

GraphFAS: A Distributed System for Automated Graph Feature Generation and Selection in Industrial Transaction Networks (arxiv.org)

The BatchNorm Illusion: Diagnosing Normalization Artifacts in Machine Unlearning Evaluation (arxiv.org)

Do Reasoning Representations Help Humans Evaluate LLM Outputs? (arxiv.org)

Earth System World Model for What-If Simulations: A Case Study for Terrestrial Ecosystems (arxiv.org)

Length Generalization for Transformers via Compression (arxiv.org)

Eliciting Weak-to-Strong Generalization with On-Policy Reverse Distillation (arxiv.org)

Adaptive Anisotropic Attention for Axis-Structured Signals (arxiv.org)

BAFF: Bid-Aware Filter Family for Mitigating Training Data Interference in RTB A/B Tests (arxiv.org)

Chimaera: A Mixture-of-Graph-Experts Architecture for Cross-Task and Cross-Dataset Graph Learning (arxiv.org)

Hyperparameter Scaling Laws Across MoE Sparsity (arxiv.org)

HOPE: Heterophily-Aware Open-Set Node Classification with Pseudo-Extrapolation (arxiv.org)

Efficient Exploration Is Enough (arxiv.org)

No-Regret Mixing of LRU and LFU with Optimal Switching Cost (arxiv.org)

CoRL: Co-Evolutionary Reinforcement Learning for Adaptive Indirect Prompt-Injection Attacks and Defenses (arxiv.org)

Statistical versus machine learning-based spatial interpolation of post-processed ensemble weather forecasts (arxiv.org)

Steering Under Compression: Dose-Response, Capability Cost, and Failure Asymmetry in Quantized LLMs (arxiv.org)

One Step, One Lead: Mitigating Higher-Order Interference in Multi-Domain Reinforcement Learning via Cross-Step Control (arxiv.org)

PAC-Bayesian Bounds for Learning Partially Observed Stochastic Linear Time-Invariant State-Space Systems with Inputs and Sub-Gaussian Noise (arxiv.org)

Local and Global Stability in Performative Reinforcement Learning (arxiv.org)

A Theoretical Framework for Masked Pretraining (MPT) (arxiv.org)

Stability and Generalization of Straight-Through Estimators for Training Two-Layer Quantized Neural Networks (arxiv.org)

Sparse Oblique Rule Boosting for Simpler Additive Rule Ensembles (arxiv.org)

Sector-Mean: Deterministic Initialization of K-Means Centroids via Angular Sector Partitioning (arxiv.org)

On BatchNorm Forward Modes in Value-Based Reinforcement Learning (arxiv.org)

MetaRSI / RSI2: A Meta-Recursive Self-Improving System for Recursive Self-Improving Systems Themselves (arxiv.org)

Are Verifier Errors Independent Within a GRPO Group? Evidence from Qwen2.5 Rollouts (arxiv.org)

Model-Adaptive and Risk-Constrained Frequency Hopping Against Predictive Jammers (arxiv.org)

Parameterized and Streaming Algorithms for Euclidean Fair $k$-Center Clustering (arxiv.org)

Robust Dynamic Expansion for Continual Learning under Backdoor Attacks via Purification and Selective Recovery (arxiv.org)

Linear Algebra Foundations of Efficient Attention: A Phase Reversal in Rank Collapse Under SVD Compression (arxiv.org)

Representation Learning for Sample-Efficient CATE Estimation by Leveraging Multiple Outcomes (arxiv.org)

SeaCausal-FL: Federated Fuzzy Causal Learning for Maritime IoT Fault Diagnosis and Counterfactual Reasoning (arxiv.org)

Robust Conformal Consensus: Multi-Agent LLM-as-a-Judge Interval Evaluation with Conformal Prediction (arxiv.org)

EgoNeMo: Transferable Map of Pedestrian Dynamics via Egocentric LiDAR Scan (arxiv.org)

Spectral Prioritized Sweeping in Nonstationary Reinforcement Learning (arxiv.org)

Decision-Aware Suffix Prediction and Reasoning of Business Processes (arxiv.org)

All for 1-Bit: Towards Genuine 1-Bit Post-Training Quantization for LLMs (arxiv.org)

Rethinking One-Shot Federated Graph Learning: Training-Free Statistical Estimation (arxiv.org)

Compressed Recurrent Feedback in Tsetlin Machines: A Reproducible Boolean-FSM Study (arxiv.org)

Sparse Incident-Cluster Learning for 12-hour Port Flood Pre-warning in Digital-Twin Analytics (arxiv.org)

DataFlex-RL: An Evaluation Platform for RLVR Data Policies (arxiv.org)

VERPO: Verified Evidence Regularized Policy Optimization (arxiv.org)

Bi-HYCO: Bi-Objective Cooperative Learning for PDE Parameter Identification under Fragmented Observations (arxiv.org)

A Computational Implementation of a Goal-Directed Theory of Affect (arxiv.org)

Evidence-Aligned Entity Verification for Hallucination Detection in Retrieval-Augmented Generation (arxiv.org)

FANS: Federated Adaptive Network Search Learning for Heterogeneous Devices (arxiv.org)

Revoked but Still Authoritative: An Empirical Study of Revocation Enforcement in Agent-Memory Systems (arxiv.org)