MemR$^3$: Memory Retrieval via Reflective Reasoning for LLM Agents

TL;DR

MemR$^3$ introduces a novel autonomous memory retrieval system for LLM agents that uses reflective reasoning and closed-loop control to improve answer quality and transparency, outperforming existing methods on key benchmarks.

Contribution

It presents MemR$^3$, a memory retrieval framework with a router and evidence tracker, enabling autonomous, transparent, and improved memory access for LLM agents.

Findings

Outperforms baselines on LoCoMo benchmark.

Improves RAG retrieval by +7.29%.

Enhances Zep retrieval by +1.94%.

Abstract

Memory systems have been designed to leverage past experiences in Large Language Model (LLM) agents. However, many deployed memory systems primarily optimize compression and storage, with comparatively less emphasis on explicit, closed-loop control of memory retrieval. From this observation, we build memory retrieval as an autonomous, accurate, and compatible agent system, named MemR$^3$, which has two core mechanisms: 1) a router that selects among retrieve, reflect, and answer actions to optimize answer quality; 2) a global evidence-gap tracker that explicitly renders the answering process transparent and tracks the evidence collection process. This design departs from the standard retrieve-then-answer pipeline by introducing a closed-loop control mechanism that enables autonomous decision-making. Empirical results on the LoCoMo benchmark demonstrate that MemR$^3$ surpasses strong baselines on LLM-as-a-Judge score, and particularly, it improves existing retrievers across four categories with an overall improvement on RAG (+7.29%) and Zep (+1.94%) using GPT-4.1-mini backend, offering a plug-and-play controller for existing memory stores.