AI论文速递 2026年06月01日（HuggingFace Daily Papers）¶

数据来源：https://huggingface.co/papers 采集时间：2026-06-01

📌 重点关注¶

Verifiable Rewards Beyond Math and Code: Lightweight Corpus-Grounded Process Supervision for Factual Question Answering | arXiv — 【重点关注】 Applying reinforcement learning to improve factual accuracy in knowledge-inte... 💡 强化学习提升知识问答准确性，突破数学与代码的验证边界
LongTraceRL: Learning Long-Context Reasoning from Search Agent Trajectories with Rubric Rewards | arXiv — 【重点关注】 Long-context reasoning remains a central challenge for large language models,... 💡 长期上下文推理新突破，直击LLM长文本处理核心痛点
CoHyDE: Iterative Co-Training of LLM Rewriter & Dense Encoder for Tool Retrieval | arXiv — 【重点关注】 Tool retrieval over large API catalogs is a core bottleneck for LLM agents: u... 💡 工具检索协同训练新范式，显著提升AI Agent执行效率

DynaFLIP: Rethinking Robotics Perception via Tri-Modal-Dynamics Guided Representation | arXiv — Robot manipulation critically depends on perception that preserves the action...
Towards Verifiable Multimodal Deep Research: A Multi-Agent Harness for Interleaved Report Generation | arXiv — Large Language Models (LLMs) have advanced autonomous agents from deep search...
Tiny but Trusted: Efficient Vision-Language Reasoning for Time-Series Anomaly Detection | arXiv — Recent advances in Vision-Language Models (VLMs) have achieved impressive per...
When Cloud Agents Meet Device Agents: Lessons from Hybrid Multi-Agent Systems | arXiv — The design space of agentic AI inference spans two extremes: frontier large l...
From Prompt Injection to Persistent Control: Defending Agentic Harness Against Trojan Backdoors | arXiv — LLM agents are evolving from conversational chatbots to operational tools in ...
Reflective Prompt Tuning through Language Model Function-Calling | arXiv — Large language models (LLMs) have become increasingly capable of following in...
Xetrieval: Mechanistically Explaining Dense Retrieval | arXiv — Explaining why dense retrievers assign high relevance scores remains challeng...