arXiv 2605.013862026-06-02cs.CL

MemORAI: 通过自适应图智能实现LLM对话代理的记忆组织与检索

MemORAI: Memory Organization and Retrieval via Adaptive Graph Intelligence for LLM Conversational Agents

Independent Researcher（独立研究者）；Hanoi University of Science and Technology（河内科学技术大学）；VNU University of Engineering and Technology（VNU工程大学）；Monash University（墨尔本大学）

Hung Pham Van, Nguyen Manh Hieu, Khang Pham Tran Tuan, Nam Le Hai, Linh Ngo Van, Nguyen Thi Ngoc Diep, Trung Le

备注：: ACL Findings

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

提出MemORAI框架，通过选择性记忆过滤、富来源追踪的多关系图存储和查询自适应子图检索，解决LLM长期对话中记忆缺失、信息稀释和检索不精准的问题，在LOCOMO和LongMemEval基准上达到最优性能。

摘要：

大型语言模型（LLM）缺乏用于长期个性化对话的持久记忆。现有的基于图的记忆系统存在信息稀释、缺乏来源追踪以及忽略查询上下文的统一检索问题。我们引入MemORAI（通过自适应图智能实现记忆组织与检索），该框架整合了三个创新：具有双层压缩的选择性记忆过滤以保留用户个性相关内容；富来源追踪的多关系图，在对话轮次级别跟踪事实来源；以及查询自适应子图检索，采用动态加权PageRank应用查询条件化的边权重。在LOCOMO和LongMemEval基准上评估，MemORAI在记忆检索和个性化响应生成方面达到了最先进的性能，表明选择性存储、富表示和自适应检索对于连贯、个性化的LLM代理至关重要。

英文摘要：

Large Language Models (LLMs) lack persistent memory for long-term personalized conversations. Existing graph-based memory systems suffer from information dilution, absent provenance tracking, and uniform retrieval that ignores query context. We introduce MemORAI (Memory Organization and Retrieval via Adaptive Graph Intelligence), a framework that integrates three innovations: selective memory filtering with dual-layer compression to retain user-persona-relevant content, a provenance-enriched multi-relational graph tracking factual origins at the turn level, and query-adaptive subgraph retrieval with Dynamic Weighted PageRank that applies query-conditioned edge weighting. Evaluated on LOCOMO and LongMemEval benchmarks, MemORAI achieves state-of-the-art performance in memory retrieval and personalized response generation, demonstrating that selective storage, enriched representation, and adaptive retrieval are essential for coherent, personalized LLM agents.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。