AI News

Learning to Defer with Guidance on Real World Medical Data (arxiv.org)

FairMean: Promoting Fairness in Distributed Learning under Label Poisoning Attacks (arxiv.org)

PACT: From Credit Assignment to Critic Alignment (arxiv.org)

SyzHarness: Patch-Based Kernel Bug Reproduction with LLM-Synthesized Fuzzing Harnesses (arxiv.org)

On the Effect of Bit-Level Parameter Perturbations in Machine Learning and Deep Learning Models (arxiv.org)

Disaggregated Quantization: Specializing LLM Prefill and Decode (arxiv.org)

PreGS: A Parameter-Transfer-Based Multi-Expert Graph Neural Network for Node Classification (arxiv.org)

CompKV: Compensation-Aware KV Selection for Long-Context LLM Inference (arxiv.org)

Quantifying Protocol-Induced Uncertainty in Comparative Predictive-Model Evaluation: Evidence from Large-Scale Daily PM10 Forecasting (arxiv.org)

Sample-Smooth Spaces: A Convenient Category for Differentiable Probabilistic Programming (arxiv.org)

FLARE: A Full-Lifecycle Dense Supervision Paradigm for Long-Horizon Coding Agents via Generative Reward Model (arxiv.org)

Team DArgk at the 2026 ELOQUENT lab for evaluating generative language model quality: Residuals of Humanity: AI Detection Evasion via GRPO Fine-Tuning (arxiv.org)

A Governance-Aware Large Language Model Orchestrated Agentic Digital Twin for Transmission System Operator Control Room Decision Support (arxiv.org)

H2LooP Telecom Model v1: From Telecom Comprehension to Autonomous Issue and PR Resolution (arxiv.org)

Damage Predicts Recovery: When Calibration Data Matters in Compressing Financial LLMs (arxiv.org)

Toolcompass: Guiding Tool Trialing, Not Suppressing It (arxiv.org)

CODA: Depth-Aligned Scene Completion and Object Decomposition from a Single RGB-D Image (arxiv.org)

Ladders of Thought: A Self-Evolving Curriculum of Progressively Simplified Reasoning Traces (arxiv.org)

Accelerating the Mitigation of LLM Inference Nondeterminism Across GPU Architectures (arxiv.org)

Evaluating Personal Information Output from Conversational Interactions in Generative AI Systems (arxiv.org)

When and How Should an Agent Clarify? CIGAsk: Teaching LLMs to Clarify via Counterfactual Information Gain (arxiv.org)

How Many Pixels Is a Digit Worth? Place-Aware Coordinate Entropy for GUI Agent Confidence Estimation (arxiv.org)

Unsupervised Brain Anomaly Detection as a Bayesian Inverse Problem with Diffusion Prior (arxiv.org)

Taming CoT Obfuscation in VLMs: From Mechanistic Evidence to Activation Enforcement (arxiv.org)

CivBench: Progress-Based Evaluation for LLMs' Strategic Decision-Making in Civilization V (arxiv.org)

HABILIS Brain 0: Geometry-Change Supervision for Vision-Language-Action and Residual Flow Recovery (arxiv.org)

Synthesis and editing of multi-instrument audio mixtures using scalar-quantised latents with MIDI Span conditioning (arxiv.org)

Matryoshka attribution: Learning to attribute language model outputs to representations and weights (arxiv.org)

FAST-ML: A Hybrid Physics-Machine Learning Framework for Tropical Cyclone Intensity Forecasting (arxiv.org)

SAM-V: Geometry-Aware Segment Anything for Multi-View Instance Segmentation (arxiv.org)

LIMIT: Less Is More for Instruction Tuning in Text-to-SQL (arxiv.org)

CREDO: Variance-Guided Rubric Evolution for Replay-Corrected Credit Assignment (arxiv.org)

APEXA: Execution-Integrity Enforcement for Multi-Agent LLM Automation of Synchrotron Data Reduction (arxiv.org)

Self-Healing Harness for Runtime Oversight of Agent Self-Modification (arxiv.org)

EDGEGEN: Improving Tool-Calling Agents Beyond Happy Paths with Synthetic Edge Case Generation (arxiv.org)

When More Evidence Hurts: Publication-Bias Drift and Principled Stopping for Biomedical Causal Search (arxiv.org)

An Implant-to-Wearable IR-UWB Transmitter-Receiver Architecture and Layered Protocol for High-Density Brain-Computer Interfaces (arxiv.org)

MIND the Gap: A Geographic Implicit Neural Representation with Adjustable Spatial Scale (arxiv.org)

WeightBridge: An Efficient Weight Transfer Library for Reinforcement Learning (arxiv.org)

From Offline Proxies to Online Decisions: A Layered Engagement Evaluation Framework for Conversational AI (arxiv.org)

Sex Estimation from Footwear Outsole Impressions Using CNN Transfer Learning and Interpretable Image Statistics (arxiv.org)

Recovering Lost Details: Multi-Scale Frequency Compensation for Long-Term Time Series Forecasting (arxiv.org)

DocMIDE: Learning Multi-Hop Implicit Derivation in Visually Rich Documents (arxiv.org)

Incremental Consistency Execution for Autonomous Intelligent Systems (arxiv.org)

Representation-guided in-context learning for medical image interpretation with multimodal large language models (arxiv.org)

Structured Decomposition for Reliable LLM-Generated Access Control Policies (arxiv.org)

Synthesizing Reactive Character Behaviors for Continuous Games via Programmatic Policy Search (arxiv.org)

Context-Aware Pre-Deployment Evaluation of AI Systems: A Regulatory Framework for Nigerian Fintech (arxiv.org)

Testing, not presuming, adequacy: calibrating generative social simulators against emergent network structure (arxiv.org)

FinInteract: Benchmarking Clarification and Intent Integration in Ambiguous Financial Question Answering (arxiv.org)