AI News

LSREP: A Longitudinal State-Replay Protocol for Evaluating Conversational Memory, with ICE v2 as an Audited Local-First Architecture (arxiv.org)

VideoMM: Adaptive Macro-Micro Inference for Efficient Video MLLMs (arxiv.org)

little m: An AI Agent for Industrial Process Optimization (arxiv.org)

AI for Games in the Foundation Model Era (arxiv.org)

ANIMASK: What the Model Contributes to Role Play in Simulated Story Worlds (arxiv.org)

ReDraft, Don't Just Distill: Reference-Driven Revision for Continual VLLM Post-Training (arxiv.org)

EchoPath: Execution-Level Replayable Memory for GUI Agents (arxiv.org)

EVINCE: Optimizing Multi-LLM Dialogues Using Conditional Statistics and Information Theory (arxiv.org)

A Framework for Generating Valid Context-Specific Benchmarks through Expert Guidance (arxiv.org)

Do LLMs Have Values? A Quantitative Analysis and Alignment Framework for Values in Large Language Models (arxiv.org)

Query-Aware Source-Risk Triage for Retrieval-Augmented Generation (arxiv.org)

QueryFormer: Winning Solution for KDD Cup 2026 Tencent UniRec Challenge (arxiv.org)

AquiLLM: Evaluating Faithfulness in Open-Weight RAG-LLM Systems for Scientific Research (arxiv.org)

Skill-based Agentic Evaluation for Real-time Data Science Tasks (arxiv.org)

Fine-Tuning Fixes Mode Collapse and Over-Dispersion in LLMs (arxiv.org)

Breaking the 1.58-bit Barrier for Ternary LLMs (arxiv.org)

Cross-Anatomy Transfer Versus Sparse Interpolation in Digital-Twin-Oriented Aortic Fluid-Structure Interaction Surrogates (arxiv.org)

BLINDSPOT: A Benchmark for Safety and Refusal Calibration in Long-Horizon Tool-Using Agents (arxiv.org)

The AI-Enabled Scientific Frontier (arxiv.org)

CADWorld: Computer-Use Benchmark for Long-Horizon Computer-Aided Design (arxiv.org)

The Pain Axis: LLMs Represent Self-Directed Harm and Act to Relieve It (arxiv.org)

Metacognitive Steering: Learning the Structure of Scientific Judgment (arxiv.org)

Toward Governance-Aware Autonomous GIS: A Narrative Review of Ethical and Privacy Risks in LLM-Enabled GeoAI (arxiv.org)

Where Should the KV Cache Live? Placement Policies Across GPU, CPU, and SSD for Long-Lived Sessions (arxiv.org)

Artificial intelligence and biosecurity: capabilities, threat pathways, and defense-in-depth governance (arxiv.org)

Calibrate, Then Route: A Measured Study of Learned Request Routing for Disaggregated LLM Serving (arxiv.org)

R3: Robust Rubric-Agnostic Reward Models (arxiv.org)

Position: AI Is Not Ready for Strategic Conflicts (arxiv.org)

GPEvac: GNN-Based PPO for Adaptive Evacuation Routing During Shooting Events (arxiv.org)

CLEAR: Cross-Source Evidence Adjudication for Large Language Models in Medicine (arxiv.org)

Safe Error Correction for Language Models: Frozen-Base Adjustment with Capability Preservation (arxiv.org)

Optimal Pruning for Neural Architectures using Fisher Information Distances (arxiv.org)

Training Energy-Based Models with Non-MCMC Samplers and Efficient Temperature Estimation (arxiv.org)

Nonnegative matrix factorizations and related compositional models: Equivalence, identifiability, and an application on the grain-size analysis of sediments (arxiv.org)

Hybrid Variational Quantum Circuits for Multivariate Regression and High-Dimensional Data Reconstruction (arxiv.org)

Conformal Policy Learning with Distribution-Free Safety Guarantees (arxiv.org)

Personalized Federated Learning through Global Knowledge Distillation and Local Head Adaptation (arxiv.org)

Covariate Selection for Doubly Robust Double/debiased Machine Learning Estimators for Causal Inference (arxiv.org)

Learning Choice Model Trees for Feature-Based Multi-Product Pricing: Exact Optimization and Field Evidence (arxiv.org)

Coupled Calibration and Learning: Mitigating Teacher Bias in LLM Distillation without Target-Domain Reward Feedback (arxiv.org)

Causal Discovery via Transformed Low-Rank Quantile Surfaces (arxiv.org)

HyCoSeq: Contextual Hyperbolic Representation Learning for Genomic Sequences (arxiv.org)

Characterizing Heterogeneous Rates in Finite Mixture Estimation via Partial Optimal Transport (arxiv.org)

Stable by Construction: Variational Latent Markov Operators for Long-Horizon PDE Prediction (arxiv.org)

Supervising the Chain Ladder (arxiv.org)

Causal Path Analysis from Perturbational and Population-Scale Single-Cell Data with Multiscale Confounding and Measurement Error (arxiv.org)

Bounded Adjustment with Reliability-Guided Embedding for Imbalanced Learning with Noisy Labels (arxiv.org)

Skeletal Prototypes on Iterative Nerve Expansions (arxiv.org)

Splitting the Difference: Interpretable Causal Forests for Treatment Effect Heterogeneity and Bias (arxiv.org)

Certified Inference and Training for Deep Equilibrium Networks: A Continuation Framework with Polynomial Complexity Guarantees (arxiv.org)