arXiv · Multiagent Systems· Zi Wang, Xingqiao Wang, Emmanuel Addai, Devika Ambekar, Xiaowei Xu·· 4 hr agoAI score36
长期记忆智能体检索准入验证框架：RHELM 与 MemOps 重分析
The Right Memory in the Wrong Context: Verifying Retrieval Admissibility in Long-Term Agent Memory
AI brief
论文提出检索准入验证框架，将每个记忆-查询对判定为 admissible、inadmissible 或 unresolved，并在匹配 required-evidence recall 下比较路由。
Source: arXiv · Multiagent Systems · arxiv.org