HeLa-Mem: Hebbian Learning and Associative Memory for LLM Agents

TL;DR

HeLa-Mem introduces a bio-inspired memory system for LLMs that models associative memory as a dynamic graph with Hebbian learning, improving long-term coherence and efficiency.

Contribution

It presents a novel memory architecture combining graph-based associative memory with Hebbian learning, inspired by cognitive neuroscience, for enhanced LLM long-term memory.

Findings

HeLa-Mem outperforms existing methods on LoCoMo across four question categories.

Uses fewer context tokens to achieve superior performance.

Employs a dual-level memory organization with episodic and semantic components.

Abstract

Long-term memory is a critical challenge for Large Language Model agents, as fixed context windows cannot preserve coherence across extended interactions. Existing memory systems represent conversation history as unstructured embedding vectors, retrieving information through semantic similarity. This paradigm fails to capture the associative structure of human memory, wherein related experiences progressively strengthen interconnections through repeated co-activation. Inspired by cognitive neuroscience, we identify three mechanisms central to biological memory: association, consolidation, and spreading activation, which remain largely absent in current research. To bridge this gap, we propose HeLa-Mem, a bio-inspired memory architecture that models memory as a dynamic graph with Hebbian learning dynamics. HeLa-Mem employs a dual-level organization: (1) an episodic memory graph that evolves through co-activation patterns, and (2) a semantic memory store populated via Hebbian Distillation, wherein a Reflective Agent identifies densely connected memory hubs and distills them into structured, reusable semantic knowledge. This dual-path design leverages both semantic similarity and learned associations, mirroring the episodic-semantic distinction in human cognition. Experiments on LoCoMo demonstrate superior performance across four question categories while using significantly fewer context tokens. Code is available on GitHub: https://github.com/ReinerBRO/HeLa-Mem