Learning to Communicate: Toward End-to-End Optimization of Multi-Agent Language Systems

TL;DR

This paper introduces DiffMAS, a framework that optimizes latent communication in multi-agent language systems, leading to improved reasoning accuracy across various benchmarks.

Contribution

It presents a novel training method for jointly learning communication encoding and interpretation in multi-agent systems using latent trajectories.

Findings

Achieves 26.7% on AIME24 reasoning benchmark.

Improves reasoning accuracy and decoding stability.

Outperforms single-agent and prior latent communication methods.

Abstract

Multi-agent systems built on large language models have shown strong performance on complex reasoning tasks, yet most work focuses on agent roles and orchestration while treating inter-agent communication as a fixed interface. Latent communication through internal representations such as key-value caches offers a promising alternative to text-based protocols, but existing approaches do not jointly optimize communication with multi-agent reasoning. Therefore we propose DiffMAS, a training framework that treats latent communication as a learnable component of multi-agent systems. DiffMAS performs parameter-efficient supervised training over multi-agent latent trajectories, enabling agents to jointly learn how information should be encoded and interpreted across interactions. Experiments on mathematical reasoning, scientific QA, code generation, and commonsense benchmarks show that DiffMAS consistently improves reasoning accuracy and decoding stability over single-agent inference, text-based multi-agent systems, and prior latent communication methods, achieving 26.7% on AIME24, 20.2% on GPQA-Diamond, and consistent gains across reasoning benchmarks.