AI News

The Physics of Multi-Turn Long-Horizon Planning: From Pre-training to Post-training via Single- and Multi-Teacher On-Policy Agentic Distillation (arxiv.org)

Multivariate Time Series Forecasting with Adaptive Non-Local Observables (arxiv.org)

Patient-Agnostic Synthetic Pretraining for Efficient Patient-Specific Intraoperative 2D/3D Registration (arxiv.org)

Similarity Is Not Logic: Factored Inference for Dual-Encoder Vision-Language Models (arxiv.org)

IndicTalk: A Large-Scale Persona-Based Multilingual Conversational Corpus for Indic Languages (arxiv.org)

SPRKD: Effective Knowledge Distillation for Deep Neural Networks via Saddle Region Approximation (arxiv.org)

Reality Monitoring in Large Language Models: Self-Knowledge That Transforms with Conversation Memory (arxiv.org)

Hybrid Semantic and Spectral Ensemble for Robust Synthetic Image Source Attribution (arxiv.org)

Calibrated Tree-Neural Fusion for Fine-Grained Vegetation Community Classification (arxiv.org)

An Empirical Study of Feature Selection Granularity (arxiv.org)

Data-Driven Diffusion Processes on Differential Forms via the Projected Ambient Connection Laplacian (arxiv.org)

On AI Safety and Security Technical Debt in Engineering AI-Enabled Systems (arxiv.org)

StanceBench: A Benchmark for Audio LLM-Based Interpersonal Stance Evaluation from Speech (arxiv.org)

Do Visual Features Improve Other-Initiated Repair Detection? A Dyadic Multimodal Approach (arxiv.org)

Two-Timescale Hierarchical Reinforcement Learning for Resilient Operations (arxiv.org)

Robust Conformalized Selection with Noisy Responses (arxiv.org)

Impute On-Demand: Adaptive Correlated Time Series Imputation for Changing Environments (arxiv.org)

Certified Parallel-in-Time Sinkhorn for Dynamic Entropic Optimal Transport (arxiv.org)

Neural Network-Driven Volatility Drag Mitigation under Aggressive Leverage (arxiv.org)

Beyond Squared Error: Exploring Loss Design for Enhanced Training of Generative Flow Networks (arxiv.org)

Token-Region Guided Cross-Attention Fusion for Multimodal Affect Interpretation (arxiv.org)

ACM: Agentic Context Management for Long Horizon Tasks (arxiv.org)

Structure Over Scale: Schema-Constrained Causal Graphs for RAG (arxiv.org)

Novel Claim or D\'ej\`a Vu? Rethinking "Contamination-Free'' Dynamic Evaluation for Multimodal Automated Fact-Checking (arxiv.org)

Scaling GUI Agents with Visual State Transitions (arxiv.org)

Reasoning or Memorization: Can LLMs Understand and Generate Chinese Xiehouyu Riddles? (arxiv.org)

Sparse Gaussian-Mixture-Model Q-Functions via Hadamard Overparametrization for Online Reinforcement Learning (arxiv.org)

EventOD: Event-Aware OD Flow Generation via LLM-Guided Semantic Modulation (arxiv.org)

Variance-Preserving Orthogonal Selection (VPOS): Greedy Feature Selection via Orthogonal Deflation in PCA Loading Space (arxiv.org)

Do Coverage and Mutation Scores of LLM-Generated Test Suites Correlate with Their Effectiveness? (Replicability Study) (arxiv.org)

Quantum-Inspired Evolutionary Neighborhood Search for Arrival-Departure Track Utilization Adjustment under Short-Term Disturbances (arxiv.org)

Diffusion-Guided Search via Exponential Tilting (DiffTilt): An Application to Falsification of Safety-Critical Systems (arxiv.org)

Analyzing the Importance of Blank for CTC-Based Knowledge Distillation (arxiv.org)

Evaluating and Mitigating the Misguidance Effect of Buggy Code in LLM-Generated Unit Tests (arxiv.org)

XGRVFL-MV: Residual-Coupled Graph-Embedded Multi-View Random Vector Functional Link Network with FleXi Guardian Loss (arxiv.org)

GeoDecider: An Evidence-Grounded Agent for Geological Interpretation via Deliberative Reasoning (arxiv.org)

Realizing Scaling Laws in Recommender Systems: A Foundation-Expert Paradigm for Hyperscale Model Deployment (arxiv.org)

Nanbeige4.2-3B: Unlocking Agentic Capabilities in a Compact Model (arxiv.org)

Latent-LoRA: Compact Latent-Space Adapters with Gradient-Free Routing for Continual Learning (arxiv.org)

The Half-Lives of Generative-AI Evidence: A 40-Record Audit, a Claim-Currency Framework, and a Reflexive Case of Frontier-Model-Assisted Research (arxiv.org)

In-Context Learning as Implicit Policy Gradient (arxiv.org)

Where Is the Cost of Third-Party API Routers in Agentic Software Development? (arxiv.org)

Monitoring Post-Disaster Urban Recovery Using High-Resolution SAR Time Series and Unsupervised Learning: Evidence from the 2023 T\"urkiye-Syria Earthquake (arxiv.org)

KAP: Bridging the Knowledge Selection-Runtime Consumption Gap in LLM Systems (arxiv.org)

Not All LLM Reasoning is Visible in the Chain-of-Thought (arxiv.org)

TokenMem: Faithful Knowledge Injection for Frozen LLMs (arxiv.org)

Scale Weight Decay and Train Better (arxiv.org)

Do Small Models Use the Law You Give Them? Context-Injected Fine-Tuning for Legal QA in Bangladesh (arxiv.org)

Why does Greedy Search produce Optimal Clustering Outcomes? A Fixed-Core Assignment Theory (arxiv.org)

Kimi K3: Open Frontier Intelligence (arxiv.org)