AI News

TreeFlash: Parallel AR-Approximation for Faster Speculative Decoding (arxiv.org)

Identifying Quantum Structure in AI Language: Evidence for Evolutionary Convergence of Human and Artificial Cognition (arxiv.org)

Edge-Aware and Content-Adaptive Infrared Gas Leak Detection for Industrial Safety Monitoring (arxiv.org)

Pretraining Language Models on Historical Text (arxiv.org)

COD10K-C: Benchmarking Robustness of Camouflaged Object Detection Under Natural Image Corruptions (arxiv.org)

Echelon: Auditable Aggregate-Only Language-Model Adaptation Across Privacy Boundaries (arxiv.org)

Multi-Modal Machine Learning for Breast Cancer Recurrence Prediction (arxiv.org)

E2LLM: Towards Efficient LLM Serving in Heterogeneous Edge/Fog Environments (arxiv.org)

Towards Compact Autonomous Driving Perception with Balanced Learning and Multi-sensor Fusion (arxiv.org)

LiveBand: Live Accompaniment Generation in the Audio Domain (arxiv.org)

Adaptive Latent Agentic Reasoning (arxiv.org)

MAdam: Metric-Aware Multi-Objective Adam (arxiv.org)

Phantom Transfer: Data Poisoning can Survive Data-Level Defences (arxiv.org)

ASAP: Exploiting the Satisficing Generalization Edge in Neural Combinatorial Optimization (arxiv.org)

q0: Primitives for Hyper-Epoch Pretraining (arxiv.org)

Finding Needles in the Haystack: Transductive Active Labeling in Ecology (arxiv.org)

IdEst: Assessing Self-Supervised Learning Representations via Intrinsic Dimension (arxiv.org)

SeSE: Black-Box Uncertainty Quantification for Large Language Models Based on Structural Information Theory (arxiv.org)

Are Common Substructures Transferable? Riemannian Graph Foundation Model with Neural Vector Bundles (arxiv.org)

DiffUNet^2: Bidirectional Prediction, Probabilistic Generation and Collaborative Visual Discovery for Scientific Data (arxiv.org)

A Geometric Lens on Physics-Aligned Data Compression (arxiv.org)

Learning Temporal Causal Structure via Smooth Differentiable Optimization (arxiv.org)

Safety Measurements for Fine-tuned LLMs Should be Grounded in Capability (arxiv.org)

MLSkip: Data Skipping for ML Filters via Lightweight Metadata (arxiv.org)

Easy-to-Use Shielding for Reinforcement Learning (arxiv.org)

Neural Navigation Functions for Zero-Shot Generalizable Motion Planning (arxiv.org)

Efficient Transformer-Based Localized Patch Sampling for Choroid Plexus Segmentation in Multiple Sclerosis (arxiv.org)

SkillPyramid: A Hierarchical Skill Consolidation Framework for Self-Evolving Agents (arxiv.org)

Plan, Verify and Fill: A Structured Parallel Decoding Approach for Diffusion Language Models (arxiv.org)

PyraMathBench: Evaluating and Improving Mathematical Capability in Large Language Models (arxiv.org)

Spectral Asymptotics of Neural Network Loss Landscapes: An Exact Decomposition of the Curvature Exponent (arxiv.org)

Testing the Test: Score-Direction Instability in Class-Split Anomaly Detection (arxiv.org)

Learning Self-Interpretation from Interpretability Artifacts: Training Lightweight Adapters on Vector-Label Pairs (arxiv.org)

Value-Aware Stochastic KV Cache Eviction for Reasoning Models (arxiv.org)

Echo-POSED: Geometric Self-Distillation for Echocardiography Guidance (arxiv.org)

AUGUSTE: Online-Learning dApp for Predictive URLLC Scheduling (arxiv.org)

Beyond Gradient Descent: Adam for Analog Ising Machines (arxiv.org)

Geometry-Aware Tabular Diffusion (arxiv.org)

ReLoRA: Knowledge-Reusing Adaptation for Fast Rollout of Evolving LLM Services (arxiv.org)

WRIT: Write-Read Intensive Trajectory Synthesis for Multi-Turn User-Facing Agents (arxiv.org)

Pruning Deep Neural Networks via the Marchenko--Pastur Distribution (arxiv.org)

Learning to See via Epiretinal Implant Stimulation in silico with Model-Based Deep Reinforcement Learning (arxiv.org)

Let There Be Light: Reflection, Refraction and Scattering for Neural Operators (arxiv.org)

Evaluating Transformer and LSTM Frameworks for Prediction in Ungauged Basins (arxiv.org)

AI Agents Enable Adaptive Computer Worms (arxiv.org)

BehaviorBench: Modeling Real-World User Decisions from Behavioral Traces (arxiv.org)

Bregman meets L\'evy: Stochastic mirror descent with heavy-tailed noise in continuous and discrete time (arxiv.org)

Hallucination Is Linearly Decodable from Mid-Layer Hidden States in Quantized LLMs (arxiv.org)

Hand Trajectory Fusion for Egocentric Natural Language Query Grounding (arxiv.org)

CL-DMDF:Dynamic Multimodal Data Fusion Model Based on Contrastive Learning (arxiv.org)