AI与计算机视觉研究最新速递（2026年4月）

人工智能(cs.AI:Artificial Intelligence)

【1】A-MAR: Agent-based Multimodal Art Retrieval for Fine-Grained Artwork Understanding

【2】A Dual Perspective on Synthetic Trajectory Generators: Utility Framework and Privacy Vulnerabilities

【3】SafetyALFRED: Evaluating Safety-Conscious Planning of Multimodal Large Language Models

【4】Time Series Augmented Generation for Financial Applications

【5】AblateCell: A Reproduce-then-Ablate Agent for Virtual Cell Repositories

【6】Multi-modal Reasoning with LLMs for Visual Semantic Arithmetic

【7】Detecting Data Contamination in Large Language Models

【8】Enhancing Construction Worker Safety in Extreme Heat: A Machine Learning Approach Utilizing Wearable Technology for Predictive Health Analytics

【9】DT2IT-MRM: Debiased Preference Construction and Iterative Training for Multimodal Reward Modeling

【10】Integrating Anomaly Detection into Agentic AI for Proactive Risk Management in Human Activity

计算机视觉与模式识别(cs.CV:Computer Vision and Pattern Recognition)

【1】Tstars-Tryon 1.0: Robust and Realistic Virtual Try-On for Diverse Fashion Items

【2】AnyRecon: Arbitrary-View 3D Reconstruction with Video Diffusion Model

【3】CityRAG: Stepping Into a City via Spatially-Grounded Video Generation

【4】Generative Drifting for Conditional Medical Image Generation

【5】ReImagine: Rethinking Controllable High-Quality Human Video Generation via Image-First Synthesis

【6】A Network-Aware Evaluation of Distributed Energy Resource Control in Smart Distribution Systems

【7】SpanVLA: Efficient Action Bridging and Learning from Negative-Recovery Samples for Vision-Language-Action Model

【8】Face Anything: 4D Face Reconstruction from Any Image Sequence

【9】Unveiling Fine-Grained Visual Traces: Evaluating Multimodal Interleaved Reasoning Chains in Multimodal STEM Tasks

【10】IR-Flow: Bridging Discriminative and Generative Image Restoration via Rectified Flow

机器学习(cs.LG:Machine Learning)

【1】Generalization at the Edge of Stability

【2】FASTER: Value-Guided Sampling for Fast RL

【3】Benign Overfitting in Adversarial Training for Vision Transformers

【4】Adaptive MSD-Splitting: Enhancing C4.5 and Random Forests for Skewed Continuous Attributes

【5】SAGE: Training-Free Semantic Evidence Composition for Edge-Cloud Inference under Hard Uplink Budgets

【6】Lyapunov-Certified Direct Switching Theory for Q-Learning

【7】Revisiting RaBitQ and TurboQuant: A Symmetric Comparison of Methods, Theory, and Experiments

【8】When Graph Structure Becomes a Liability: A Critical Re-Evaluation of Graph Neural Networks for Bitcoin Fraud Detection under Temporal Distribution Shift

【9】EVPO: Explained Variance Policy Optimization for Adaptive Critic Utilization in LLM Post-Training

【10】Revisiting Catastrophic Forgetting in Continual Knowledge Graph Embedding

图像与视频处理(eess.IV:Image and Video Processing)

【1】Deep Image Prior for photoacoustic tomography can mitigate limited-view artifacts

【2】A Controlled Benchmark of Visual State-Space Backbones with Domain-Shift and Boundary Analysis for Remote-Sensing Segmentation