AI News

DCRL: Decoupling and Coupling Reinforcement Learning via Policy-Reward Manifold Alignment (arxiv.org)

VCMM: Variance-Calibrated Momentum for Multimodal Learning (arxiv.org)

Does Step Law Transfer to Small-Scale Language Models? An Empirical Recalibration Below 59M Parameters (arxiv.org)

The Capability Manifold and ML Scaling Laws (arxiv.org)

Hidden not Deleted: How Networks Suppress Entangled Features (arxiv.org)

MicroQonv: Reshaping Convolution Tensors for Efficient Microscaling in Training and Inference (arxiv.org)

AnchorReasoning: A Visual Grounding and Causal Reasoning Dataset in Long-Tail Autonomous Driving Scenarios (arxiv.org)

Pheno-GS: Phenoscape-scale Geodesic Sinkhorn (arxiv.org)

Learning Local Heterogeneity and Cross-Region Context for Large-Scale Traffic Forecasting (arxiv.org)

FLEET: From Logits Entropy to Enhanced Trajectories in Text Generation (arxiv.org)

Can LLMs Reason About Runtime Behavior? A Repository-Level Dynamic Benchmark (arxiv.org)

Where Should I Join? Robot Group Joining via Language-Guided Goal Prediction (arxiv.org)

Retrieval Augmented (Knowledge Graph), and Large Language Model-Driven Design Structure Matrix (DSM) Generation of Cyber-Physical Systems (arxiv.org)

The Tasteful Agent: Measuring and Improving Taste in Long-Horizon Tasks (arxiv.org)

A Scalable Multi-Robot Framework for Decentralized and Asynchronous Perception-Action-Communication Loops (arxiv.org)

Small Cues, Big Consequences: Learning Pivotal Cues for Multimodal Meme Classification (arxiv.org)

Noise-Induced Predictability Redistribution Across Forecast Horizons of Extreme Events in Chaotic Dynamics (arxiv.org)

Robust Adversarial Reinforcement Learning with Risk Sensitivity and Critic Consistency Regularization (arxiv.org)

What Do Tabular Foundation Models Compute In Context? In-Situ Representation Refinement through Attention-Gated Updates (arxiv.org)

Limiting-Kernel Q($\lambda$): Bridging Short and Long Horizons (arxiv.org)

Backdoors Leave Structural Traces: FedMAST for Backdoor Detection and Containment in Federated Learning (arxiv.org)

Learning to Detect Symbolic Failure: Machine Learning and the Limits of Black-Scholes (arxiv.org)

When Adaptation Hurts: Split Sensitivity and Person-Level Negative Transfer in Federated Wearable Onboarding (arxiv.org)

CAST: Context- and Anomaly Structure-Conditioned Time Series Anomaly Generation (arxiv.org)

From Reasoning Strings to Partial Orders: Verifier-Certified Rule Transport through Quotient Policy Optimization (arxiv.org)

Exact Minimax One-Bit Unbiased Compression: Heavy-Tail Necessity and Finite-Randomness Approximation (arxiv.org)

What Changed? Drift Detection with Real, Virtual, and Incomparable Diagnosis (arxiv.org)

ChronosAttack: Adversarial Tool Scheduling Attacks on LLM Agents (arxiv.org)

Query Implied Generative Engine Optimization (arxiv.org)

A Shared Encoder Is Not a Shared Task: Conditional Comparison for Deep Expert Pools (arxiv.org)

Evaluation Choices Decide the Forecasting Leaderboard: Evidence from a Production Marketplace Panel (arxiv.org)

False-science induction in autonomous scientific discovery (arxiv.org)

Global tree forecasters collapse at the hierarchical aggregate: a five-panel failure characterization (arxiv.org)

Spread and Scale: What Determines Whether Test-Time Budget Allocation Pays (arxiv.org)

Alignment of LRMs via Counter-Aligned Few-Shot Conversation Exposure (arxiv.org)

Binary Quantized Neural Network Training Is W[1]-Hard Parameterized by Input and Output Dimensions (arxiv.org)

Quality over Quantity: Semi-Supervised Detection of Illicit Bitcoin Flows via Feature Engineering (arxiv.org)

When Accuracy Gaps Fail to Certify: Auditing Cross-Domain Recalibration of LLM Judges (arxiv.org)

Semantic Self-Distillation for Language Model Uncertainty (arxiv.org)

CS-WCP: Robust Conformal Sets for LLM-Judge Traffic Shifts with Uncertain Group Proportions (arxiv.org)

"What's That Sound?": A Versatile, Robust, and Lightweight Convolutional Transformer for Environment Sound Recognition (arxiv.org)

I-SplineFlow: Learning Monotone Spline Stochastic Interpolant Schedulers for Few-Step Generation (arxiv.org)

Linear RNN Scaling Laws: When Longer Sequences Beat More Sequences (arxiv.org)

Riemannian Structure and Optimization for a Class of Low-Parametric Orthogonal Matrices (arxiv.org)

From Sentiment Classification to Actionable and Responsible Feedback: A Scoping Review and Evidence Map of NLP in Student Evaluation of Teaching, 2015-2026 (arxiv.org)

Agentic Governance and Adversarial Verification for Policy-Constrained LLM Healthcare Appeal Generation (arxiv.org)

Reachable Global Optimization in AI Systems: How Global Is Global? (arxiv.org)

A hierarchy of faithfulness criteria for knowledge base completion (arxiv.org)

Learning What to Activate: Combinatorial Capability Allocation for Long-Horizon Multimodal Agents (arxiv.org)

Hard Negatives Reveal What Easy Negatives Hide: Cross-Lingual Harmfulness Representations Degrade with Resource Tier Under Hard Negatives (arxiv.org)