INMS: Memory Sharing for Large Language Model based Agents

TL;DR

INMS introduces an asynchronous memory sharing framework for multi-agent LLM systems, enabling dynamic knowledge exchange and improving performance through shared conversational memory.

Contribution

The paper presents INMS, a novel framework for real-time, asynchronous memory sharing among agents, enhancing multi-agent collaboration and knowledge integration.

Findings

Significantly improves agent performance across datasets.

Effectively models multi-agent interaction and collective knowledge sharing.

Enhances dynamic knowledge exchange in large language model agents.

Abstract

While Large Language Model (LLM) based agents excel at complex tasks, their performance in open-ended scenarios is often constrained by isolated operation and reliance on static databases, missing the dynamic knowledge exchange of human dialogue. To bridge this gap, we propose the INteractive Memory Sharing (INMS) framework, an asynchronous interaction paradigm for multi-agent systems. By integrating real-time memory filtering, storage, and retrieval, INMS establishes a shared conversational memory pool. This enables continuous, dialogue-like memory sharing among agents, promoting collective self-enhancement and dynamically refining the retrieval mediator based on interaction history. Extensive experiments across three datasets demonstrate that INMS significantly improves agent performance by effectively modeling multi-agent interaction and collective knowledge sharing.