Multi-Agent Transactive Memory

2026-06-18

AI Summary

Existing heterogeneous large language model agents lack mechanisms for cross-agent knowledge sharing, leading to redundant exploration. This work proposes the first population-level trajectory storage and retrieval framework that enables experience reuse through shared memory, without requiring joint training or explicit coordination. Built upon retrieval-augmented generation, the framework constructs a shared trajectory repository to efficiently retrieve long-horizon procedural knowledge in environments such as ALFWorld and WebArena. Experimental results demonstrate that the approach significantly improves task success rates and reduces the number of interaction steps, thereby validating the effectiveness and scalability of population-level experience sharing.

Abstract

The decentralized deployment of LLM agents with diverse capabilities across diverse tasks motivates infrastructure for knowledge sharing across heterogeneous agent populations. Just as search engines index human-generated artifacts to support human problem solving, retrieval systems can organize agent-generated artifacts for reuse across agent populations. We extend retrieval-augmented generation - which demonstrates the value of human-authored artifacts to individual agents - to retrieval of agent-generated artifacts supporting a population of agents. In particular, agent trajectories encode reusable procedural knowledge, yet these artifacts are typically discarded after a single use or retained only by the producing agent, forcing newly instantiated agents to repeatedly rediscover existing solutions. We propose Multi-Agent Transactive Memory (MATM), a framework for population-level storage and retrieval of agent-generated trajectories, where producer agents contribute trajectories to a shared repository and consumer agents retrieve them to improve task execution. We focus on interactive environments (ALFWorld and WebArena), where trajectories are long and encode especially rich procedural structure. Our experiments demonstrate that retrieving trajectories from MATM improves downstream task performance and reduces interaction steps without coordination or joint training. These results position MATM as a design pattern for population-level experience sharing in open agent ecosystems.

Problem

• Multi-Agent Systems
• Transactive Memory
• Agent Trajectories
• Knowledge Sharing
• Retrieval-Augmented Generation

Innovation

• Multi-Agent Transactive Memory
• retrieval-augmented generation
• agent trajectories
• knowledge sharing
• heterogeneous agents