AI News

Collaborative and Efficient Fine-tuning: Leveraging Task Similarity (arxiv.org)

Universal One-third Time Scaling in Learning Peaked Distributions (arxiv.org)

Independent Component Discovery in Temporal Count Data (arxiv.org)

Safeguarded Stochastic Polyak Step Sizes for Non-smooth Optimization: Robust Performance Without Small (Sub)Gradients (arxiv.org)

Trajectory Data Suffices for Statistically Efficient Policy Evaluation in Fixed-Horizon Offline RL with Linear $q^\pi$-Realizability and Concentrability (arxiv.org)

Interpretable Self-Supervised Learning via Representer Landmarks and Nystr\"om Approximation (arxiv.org)

Human in the Loop Adaptive Optimization for Improved Time Series Forecasting (arxiv.org)

Global Convergence of Adaptive Sensing for Principal Eigenvector Estimation (arxiv.org)

Advancing Local Clustering on Graphs via Compressive Sensing: Semi-supervised and Unsupervised Methods (arxiv.org)

Non-vacuous Generalization Bounds for Deep Neural Networks without any modification to the trained models (arxiv.org)

Targeted Data Fusion for Region-Specific Survival Effects in the AMP HIV Prevention Trials (arxiv.org)

Towards Simple and Provable Parameter-Free Adaptive Gradient Methods (arxiv.org)

Challenges in the calibration of tree-based models for imbalanced classification (arxiv.org)

Fixed-Mean Gaussian Processes for Post-hoc Bayesian Deep Learning (arxiv.org)

Dimension Reduction via Sum-of-Squares and Improved Clustering Algorithms for Non-Spherical Mixtures (arxiv.org)

A Likelihood Approach for Inference of Population Heterogeneity in Particle Ensembles with Second-Order Langevin Dynamics (arxiv.org)

Near-Optimal and Tractable Estimation under Shift-Invariance (arxiv.org)

Optimizing accuracy and diversity: a multi-task approach to forecast combinations (arxiv.org)

Adaptive Querying with AI Persona Priors (arxiv.org)

Beyond Procedure: Substantive Fairness in Conformal Prediction (arxiv.org)

Rethinking Bregman Divergences in Kronecker-Factored Optimizers (arxiv.org)

PropLLM: Propagation-Aware Scene Reconstruction for Network Fault Diagnosis (arxiv.org)

HomeFlow: A Data Flywheel for Smart Home Agent Training with Verifiable Simulation (arxiv.org)

Application of Algorithms in Energy-Efficient Design Platforms for Green Building (arxiv.org)

Can LLM Agents Sustain Long-Horizon Organizational Dynamics? (arxiv.org)

The Case for Model Science: Verify, Explore, Steer, Refine (arxiv.org)

"Skill issues'': data-centric optimization of lakehouse agents (arxiv.org)

Deft Scheduling of Dynamic Cloud Workflows with Varying Deadlines via Mixture-of-Experts (arxiv.org)

Emergent Ordinal Geometry in Transformers Trained on Local Comparisons (arxiv.org)

DREAM-S: Speculative Decoding with Searchable Drafting and Target-Aware Refinement for Multimodal Generation (arxiv.org)

Generate in Reconstruction Space, Match in Semantic Space: Transport Geometry for One-Step Generation (arxiv.org)

Saliency-Aware Model Merging (arxiv.org)

TabChange: Precise Attribute Changes in Tabular Data (arxiv.org)

Torus Graphs for Large Scale Neural Phase Analysis (arxiv.org)

Expected Value Alignment for Generative Reward Modeling in Formal Mathematics Verification (arxiv.org)

Reasoning4Sciences: Bridging Reasoning Language Models to All Scientific Branches (arxiv.org)

SkillRevise: Improving LLM-Authored Agent Skills via Trace-Conditioned Skill Revision (arxiv.org)

Diagnosing LLM Arbitration Behavior over Pre-evidence Epistemic States in RAG-based Fact-Checking (arxiv.org)

CAREAgent: Clinical Agent with Structured Reasoning and Tool-Integrated for Order Generation (arxiv.org)

ProjQ: Project-and-Quantize for Adapter-Aware LLM Compression (arxiv.org)

EST-PRM: Stress-Testing Process Reward Models Before They Become Load-Bearing (arxiv.org)

Grounded Decoding: Retrieval-Anchored Probability Fusion for Faithful RAG (arxiv.org)

Variance-sensitive Thompson sampling for generalised linear bandits, revisited (arxiv.org)

Topology-Aware State Abstraction with Tangle Cores for Markov Decision Processes (arxiv.org)

Deep Learning as the Disciplined Construction of Tame Objects (arxiv.org)

Before the Model Learns the Bug:Fuzzing RLVR Verifiers (arxiv.org)

MindClaw: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention (arxiv.org)

DAG-MoE: From Simple Mixture to Structural Aggregation in Mixture-of-Experts (arxiv.org)

AnyEdit++: Adaptive Long-Form Knowledge Editing via Bayesian Surprise (arxiv.org)

TravelEval: A Comprehensive Benchmarking Framework for Evaluating LLM-Powered Travel Planning Agents (arxiv.org)