AI News

Degree-Corrected Joint Matrix Factorization for Multilayer Community Detection (arxiv.org)

Structured-Noise Masked Modeling for Video, Audio and Beyond (arxiv.org)

Generation Provenance Before Behavior Attribution: Auditing Synthetic Speech Research Objects (arxiv.org)

SupraTITO: Transferable Generative Molecular Dynamics for Supramolecular Systems (arxiv.org)

Smoother Flow Matching via Contrastive Trajectory Repulsion (arxiv.org)

Optimal Transport Meets Reinforcement Learning: A Survey (arxiv.org)

Generalization Is Stability, Not Accuracy: Multi-Axis Evaluation of LLMs (arxiv.org)

Learning ab initio phase-field models (arxiv.org)

LESS: Lightweight Evolutionary Supernet Search in Minutes (arxiv.org)

Zero Flux: Flow-Based Comparison of High-Dimensional Discrete Distributions (arxiv.org)

FedMAD: Modulation-Aware Directional Aggregation for Federated Learning in Remote Sensing Image Classification (arxiv.org)

Rethinking the Global Convergence of Softmax Policy Gradient with Linear Function Approximation: The Case of Multi-Armed Bandits (arxiv.org)

Forward Target Propagation: A Forward-Only Approach to Global Error Credit Assignment via Local Losses (arxiv.org)

The Benchmarking Epistemology: Validity Theory for Evaluating Machine Learning Models (arxiv.org)

Causal-Aware Tabular GANs with Reinforcement Learning (arxiv.org)

Don't Waste the Noise: Importance-Guided Perturbation Allocation under Joint Global and Local Constraints (arxiv.org)

Rethinking Data Augmentation under Covariate Shift: Invariant-Guided Diffusion and Prototype Reweighting (arxiv.org)

Match the Distribution, Not the Compute: Post-Training Multi-Token Prediction Heads (arxiv.org)

Cross-Benchmark Transfer from RL on Agentic Coding Tasks (arxiv.org)

Clock Diffusion: Efficient Semi-Autoregressive Continuous Diffusion Language Models (arxiv.org)

Towards Fast and Disentangled Counterfactuals for Visual Foundation Models (arxiv.org)

When Do Biological Reasoning Models Use Their Biological Inputs? (arxiv.org)

Rate-Optimal Algorithm for Adversarial Linear CMDPs (arxiv.org)

Platonic Task Arithmetic (arxiv.org)

GUI-HARVEST: Self-Improving GUI Agents through Evidence-Driven Harness Evolution (arxiv.org)

Structure-agnostic Causal Representation Learning (arxiv.org)

Variational Streaming Flow: Probabilistic Forecasting in Physical Time (arxiv.org)

Generalist Representation, Specialist Detection: TS-Router for Time-Series Anomaly Detection (arxiv.org)

HADRec: A Hierarchy-Aware Drug Recommendation Framework by Fusing Molecular Knowledge and Electronic Health Record (arxiv.org)

Neural scaling laws and evolution of learnable activation functions of Kolmogorov-Arnold networks (arxiv.org)

Auditable Algebraic Counting Field for Cryptic-Pocket Detection from Apo Structures (arxiv.org)

Reliability-aware short-term roll prediction for unmanned surface vehicles via multi-task learning and adaptive centralization (arxiv.org)

Optimal Transport Reweighting for Robust Learning under Spurious Correlations and Label Noise (arxiv.org)

SLIM: Simplex-Lattice Interpolation Merging (arxiv.org)

Kernelized Activation Steering (arxiv.org)

GLoC-EHR: Evidence-Cited Clinical Reasoning over Global Context and Local EHR Events (arxiv.org)

Dataset Identity, Not Novelty: The Source of an Inflated OOD Detection Gain (arxiv.org)

How Much Can Language Models Gain from Test-Time Computation? (arxiv.org)

Latent Information Sharing for Accelerating Federated Learning (arxiv.org)

Does Scaling Reinforcement Learning Really Require More Training? (arxiv.org)

Parameter-Efficient Distributionally Robust Adaptation of Tabular Foundation Models under Subpopulation Shift (arxiv.org)

Persistent Depth Ordering amid Shifting Block-Bypass Responses in Language Model Pretraining (arxiv.org)

Detect, Explain, Interpret: An End-to-End Benchmark for Time Series Anomaly Detection, Explainability and Interpretability (arxiv.org)

Learning Rate Transfer for Hybrid Transformer-SSM Architectures (arxiv.org)

CAGE-NAS: Certified Functional Descent for Efficient Model Growth (arxiv.org)

Rethinking the Information Bottleneck: Structured Decomposition under Label-Induced Partitions (arxiv.org)

Fully Online Decentralized Learning in Stochastic Games with Unknown Independent Chains (arxiv.org)

Counterfactual Generation via Flow Matching: Coupling-Sensitive End-to-End Rates (arxiv.org)

Low-Budget Active Learning through Entropic Optimal Transport (arxiv.org)

Early Learning Shapes Later Directions Of Representation Change In Continual Learning (arxiv.org)