AI News

Composing What Each Teacher Learned: Multi-Teacher On-Policy Distillation through Teacher-Relative Shifts (arxiv.org)

Distilling Graph Geometry: Knowledge Gap from GNNs to MLPs (arxiv.org)

Multi-LLM Collaborative Alignment via Stackelberg Games (arxiv.org)

Residual-Stream Burden Shapes Representation Learning in Diffusion Transformers (arxiv.org)

Decoupling Exploration from Optimization in RLVR (arxiv.org)

Deep Learning-Based Tri-Hybrid Multi-User MIMO Precoding: The Blessing of EM-Reconfigurable Antennas (arxiv.org)

HE-OFT: Privacy-Preserving One-Shot Federated Fine-Tuning under Homomorphic Encryption (arxiv.org)

Accelerating Floating-Point Satisfiability Solving via Gradient Normalization (arxiv.org)

On the Cyclic Assumption of the Cow-Path Search Algorithm (arxiv.org)

Train4Merge: A Controlled Single-Teacher Study of RL vs. SFT Teachers for OPD-Based Model Merging (arxiv.org)

Route-Verify-Vote: Procedure-Conditioned Self-Consistency for Mixed-Domain Reasoning (arxiv.org)

Child ASR Adaptation with Adult Retention: An Empirical Study (arxiv.org)

CoDR: Training-Free Confidence-Drift Remasking for Diffusion Language Models (arxiv.org)

Beyond the Sycophancy Score: How Task, Model, and Pressure Shape LLM Yielding (arxiv.org)

Walk fast but be careful: Understanding Parallel Sampling in Masked Diffusion (arxiv.org)

Beyond Risk Prediction: Evidence Grounding and Psychosocial Factor Verification for Explainable Suicide Risk Assessment (arxiv.org)

Autonomous Droplet Navigation via Model-Based Reinforcement Learning: Zero-Shot Transfer and Emergent Dynamics (arxiv.org)

LRCC: Generalizing Low-Rank Compression with Conditional Computation (arxiv.org)

Adversarial RL for Port-Scan Evasion: Attacker Feature Visibility in Edge-Deployed IDS (arxiv.org)

Democratizing MoE inference on commodity GPUs with CoMoE (arxiv.org)

DiTS: Multimodal Diffusion Transformers Are Time Series Forecasters (arxiv.org)

Geometry-Aware Diffusion Approximate Posterior Sampling for Sparse-View and Limited-Angle CT (arxiv.org)

Learned Monotone Recurrent Features in Governed Credit Scoring: The Price of the Frame and the Necessity of Macro Conditioning (arxiv.org)

An Empirical Study of Agent Skills' Downstream Utility (arxiv.org)

CATune: Structural Constraint-Aware Bayesian Optimization for DBMS Configuration Tuning (arxiv.org)

RT-Safe: Benchmarking Agent Safety in Real-Time Embodied Environment (arxiv.org)

Attention-Mass Condensation for Sparse Decoding (arxiv.org)

emg2face: Expressive Facial Animation with High-Density Surface EMG (arxiv.org)

Visual Jev Rewards: Reference-Bound Verification for Multi-Subject Image Generation (arxiv.org)

VM-ARRAYDPS: Virtual Microphone Augmented Diffusion Posterior Sampling for Unsupervised Blind Speech Separation (arxiv.org)

SearchWorld: Spatial Value-Grounded Imagination for UAV Object Search via World Models (arxiv.org)

PrismSSL: One Interface, Many Modalities; A Single-Interface Library for Multimodal Self-Supervised Learning (arxiv.org)

Decentralized SGD under Heavy-Tailed Noise: Optimal Convergence Rates and the Role of Gradient Clipping (arxiv.org)

Closing the Loop on Contrail Avoidance with Satellite Verification (arxiv.org)

Rephrase Before You Act: Characterizing and Mitigating Language Sensitivity in Vision-Language-Action Models (arxiv.org)

SciExam for ENSO: Can AI Agents Build Climate Models? (arxiv.org)

Pareto-optimal quantum kernel selection for unsupervised anomaly detection on real malware beaconing data (arxiv.org)

Boundary-aware Reinforcement Learning for Hypercube State Spaces via Deterministic Policy Gradient (arxiv.org)

Understanding and Mitigating Token-Pruning-Induced Vulnerabilities in VLMs (arxiv.org)

Optimal Regret for Online Market Making with Limit Order Book (arxiv.org)

NAViLoss: An Underwater Navigation-Aware Dual-Residual Objective for Physics-Consistent Learning (arxiv.org)

A Multi-Source Ultrasound Benchmark Revealing the Limits of Contemporary Self-Supervised Anomaly Detection Methods (arxiv.org)

Q-PhotoMarket: A Design Space Exploration Framework for Photonic Hybrid Quantum Neural Networks in Financial Market Prediction (arxiv.org)

Tracing Inputs, Verifying Outputs: Validating Attribution in Music Generation (arxiv.org)

Immiscible Diffusion Policy: Preserving Multimodal Robot Actions through Label-Free Noise Assignment (arxiv.org)

Quantum anomaly detection in real scarce data (arxiv.org)

How Do Agentic LLMs Decide to Call Tools? A Tool-Call Vector Shaped by Suppression (arxiv.org)

Second-order optimization of variable projection SVM models and road abnormality detection (arxiv.org)

The Persona Hierarchy Model: Understanding Contextual Generalization in Fine-Tuning LLMs (arxiv.org)

An Invariant Tangent-Angle Descriptor and a Band U-Net for 2D Fragment Adjacency Prediction (arxiv.org)