AI News

From Task Mixtures to Specialized Experts (arxiv.org)

AIR-LLM: Broadcasting AI Weights over Radio for Memory-Free Edge LLM Inference via RF Computing (arxiv.org)

Function-Structured Reinforcement Learning with Executable Verifiers for Mathematical Reasoning (arxiv.org)

Fixed-point neural samplers on discrete spaces (arxiv.org)

Exact Distinguishability in Non-Markovian Decision Processes (arxiv.org)

PACT: End-to-End Learning of Human Pose, Contacts, and Forces from Video (arxiv.org)

FineART: Fine-Grained Annotated Robotic Trajectory Dataset and Vision-Language-Action Model for Bimanual Manipulation (arxiv.org)

Iterative Topic Taxonomy Induction with LLMs: A Case Study of Electoral Advertising (arxiv.org)

ChainLoRA: Geometry-Preserving Task Vector Merging for Continual Learning in LLMs (arxiv.org)

Target-Dependent Limits of Causal Repair: A Leading-Log Frontier in a Gaussian Model (arxiv.org)

Towards a Cloud Fog Edge System for Smart Building (arxiv.org)

Beyond Unimodal Bases: Pullback Geometry for Multimodal Data (arxiv.org)

Geometry-Aware Adaptation for Pretrained Models (arxiv.org)

StoCFL: A Stochastically Clustered Federated Learning Framework for Non-IID Data with Dynamic Client Participation (arxiv.org)

Leto: Fast In-Place Recovery for LLM Training on Surviving Hardware (arxiv.org)

SemanTok: Predictable Semantic Tokens for Efficient Autoregressive Video Generation (arxiv.org)

SkillEvoLean: Mutation-enhanced skill evolution for Lean provers (arxiv.org)

Group-Invariant Statistics Determine Embedding Geometry: Harmonic Analysis of Representations from Bach to the Night Sky (arxiv.org)

Learning Linear Systems under Heavy-Tailed Noise: A Non-Asymptotic Analysis from A Single Trajectory (arxiv.org)

Format-Aware Fusion for Fast FP4 Pretraining (arxiv.org)

Specificity-Aware Diffusion Steering via Variance-Reduced Sequential Monte Carlo (arxiv.org)

Forking: Sudden Overfitting Under Replay (arxiv.org)

Interpretable Synthetic Medical Tabular Data Generation for Clinical Decision Support Using Fuzzy Cognitive Maps (arxiv.org)

MatrixReward: Reward from Rubric Matrix for Open-Ended Generation (arxiv.org)

UniGuardian: A Unified Defense for Detecting Prompt Injection, Backdoor Attacks and Adversarial Attacks in Large Language Models (arxiv.org)

Do Your Own Research: Learning to Forecast by Learning to Search (arxiv.org)

Sharp Non-Asymptotic Analysis of the Penalized Challenger in $\beta$-EB-TCI for Bernoulli Bandits (arxiv.org)

The Curvature of Regret in Contextual Linear Optimization (arxiv.org)

Latent JEPA: Abstract Future Prediction for Latent Reasoning in Chemistry (arxiv.org)

Token Space: A Category Theory Framework for AI Computations (arxiv.org)

Provable FDR Control for Deep Feature Selection: Deep MLPs and Beyond (arxiv.org)

A framework for auditing grounding claims (arxiv.org)

Graph Representation via Elements of Discrete Morse and Cobordism Theories (arxiv.org)

Learning to Predict Distributions over Weight Updates for Test-Time Adaptation (arxiv.org)

Same Reward, Different Skills: When Multimodal RL Learns to Look (arxiv.org)

Convergent Plug-and-Play Image Restoration with Annealed Noise Levels (arxiv.org)

BalLOT: Balanced $k$-means clustering with optimal transport (arxiv.org)

How many labelers do you have? A closer look at gold-standard labels (arxiv.org)

Fewer Tokens, More Self-Teaching: On-Policy Self-Distillation for Extreme Visual Token Reduction (arxiv.org)

Error-Corrected Inference-Time Scaling for Imperfect Diffusion Models (arxiv.org)

LAST: Looped Audio Spectrogram Transformer (arxiv.org)

A Structured State Space Sequence Model for Multi-Class Classification of Malware (arxiv.org)

TRACE: Tackling Real-World Resource Assignment Problems via Agentic Heuristic Design (arxiv.org)

Beyond Decodability: Do Acoustic Factors Drive Predictions in Speech-Based Alzheimer's Assessment? (arxiv.org)

Optimal Stochastic Bilevel Optimization with First-Order Oracles (arxiv.org)

Varda-single-1.0: deterministic data-driven weather forecasting at 1 km resolution over Switzerland's complex topography (arxiv.org)

Code Owns the Simulation, Jev Owns the Evaluation (arxiv.org)

Compressing History into Memory: Distilling Transformers into Recurrent Transformers (arxiv.org)

Deep Symmetric Autoencoders from the Eckart-Young-Schmidt Perspective (arxiv.org)

Generalized Engression Models (arxiv.org)