LiCoMemory: Lightweight and Cognitive Agentic Memory for Efficient Long-Term Reasoning

TL;DR

LiCoMemory introduces a lightweight hierarchical graph memory system for LLM agents, significantly improving long-term reasoning, retrieval efficiency, and multi-session consistency in dialogue tasks.

Contribution

The paper presents LiCoMemory, a novel hierarchical graph-based memory framework that enhances real-time knowledge updating and retrieval for LLM agents, addressing limitations of existing flat memory structures.

Findings

Outperforms baselines in temporal reasoning and multi-session consistency

Reduces update latency significantly

Enhances retrieval efficiency in long-term dialogue tasks

Abstract

Large Language Model (LLM) agents exhibit remarkable conversational and reasoning capabilities but remain constrained by limited context windows and the lack of persistent memory. Recent efforts address these limitations via external memory architectures, often employing graph-based representations, yet most adopt flat, entangled structures that intertwine semantics with topology, leading to redundant representations, unstructured retrieval, and degraded efficiency and accuracy. To resolve these issues, we propose LiCoMemory, an end-to-end agentic memory framework for real-time updating and retrieval, which introduces CogniGraph, a lightweight hierarchical graph that utilizes entities and relations as semantic indexing layers, and employs temporal and hierarchy-aware search with integrated reranking for adaptive and coherent knowledge retrieval. Experiments on long-term dialogue benchmarks, LoCoMo and LongMemEval, show that LiCoMemory not only outperforms established baselines in temporal reasoning, multi-session consistency, and retrieval efficiency, but also notably reduces update latency. Our official code and data are available at https://github.com/EverM0re/LiCoMemory.