EvolveMem:Self-Evolving Memory Architecture via AutoResearch for LLM Agents

Jiaqi Liu, Xinyu Ye, Peng Xia, Zeyu Zheng, Cihang Xie, Mingyu Ding, Huaxiu Yao

arXiv:2605.13941 · 2026-05-15 공개 · arXiv · PDF

llm-agents auto-research locomo membench closed-loop-optimization self-evolving-memory llm-powered evolve-mem

Abstract

Long-term memory is essential for LLM agents that operate across multiple sessions, yet existing memory systems treat retrieval infrastructure as fixed: stored content evolves while scoring functions, fusion strategies, and answer-generation policies remain frozen at deployment. We argue that truly adaptive memory requires co-evolution at two levels: the stored knowledge and the retrieval mechanism that queries it. We present EvolveMem, a self-evolving memory architecture that exposes its full retrieval configuration as a structured action space optimized by an LLM-powered diagnosis module. In each evolution round, the module reads per-question failure logs, identifies root causes, and proposes targeted configuration adjustments; a guarded meta-analyzer applies them with automatic revert-on-regression and explore-on-stagnation safeguards. This closed-loop self-evolution realizes an AutoResearch process: the system autonomously conducts iterative research cycles on its own architecture, replacing manual configuration tuning. Starting from a minimal baseline, the process converges autonomously, discovering effective retrieval strategies including entirely new configuration dimensions not present in the original action space. On LoCoMo, EvolveMem outperforms the strongest baseline by 25.7% relative and achieves a 78.0% relative improvement over the minimal baseline. On MemBench, EvolveMem exceeds the strongest baseline by 18.9% relative. Evolved configurations transfer across benchmarks with positive rather than catastrophic transfer, indicating that the self-evolution process captures universal retrieval principles rather than benchmark-specific heuristics. Code is available at https://github.com/aiming-lab/SimpleMem.

한국어 요약

📋 한 줄 요약

**[LLM 에이전트 · 장기 메모리]** 저장 콘텐츠뿐 아니라 검색 메커니즘 자체를 LLM 진단 모듈로 공진화시키는 자기 진화 메모리 아키텍처 EvolveMem을 제안한다.

🎯 핵심 기여도

💡 핵심 아이디어

"진정한 적응형 메모리"는 무엇을 저장할지뿐 아니라 무엇으로 검색할지도 함께 진화해야 한다. EvolveMem은 실패 로그를 LLM 진단으로 읽어 구성 조정을 제안하고 자동 안전장치로 적용한다.

🔬 기술적 접근법

📊 주요 결과

💭 의의 및 한계

**의의**: 메모리 시스템 설계 자체를 자동화하는 AutoResearch 패러다임을 실증한다. **한계**: LLM 진단 모듈에 의존하므로 진단 능력의 품질과 비용이 전체 성능에 영향을 미친다.

🚀 실용적 활용