Towards Language-Augmented Multi-Agent Deep Reinforcement Learning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Toquebiau, Maxime, Jun, Jae-Yun, Benamar, Faïz, Bredeche, Nicolas
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916884650131456
author Toquebiau, Maxime
Jun, Jae-Yun
Benamar, Faïz
Bredeche, Nicolas
author_facet Toquebiau, Maxime
Jun, Jae-Yun
Benamar, Faïz
Bredeche, Nicolas
contents Most prior works on communication in multi-agent reinforcement learning have focused on emergent communication, which often results in inefficient and non-interpretable systems. Inspired by the role of language in natural intelligence, we investigate how grounding agents in a human-defined language can improve the learning and coordination of embodied agents. We propose a framework in which agents are trained not only to act but also to produce and interpret natural language descriptions of their observations. This language-augmented learning serves a dual role: enabling efficient and interpretable communication between agents, and guiding representation learning. We demonstrate that language-augmented agents outperform emergent communication baselines across various tasks. Our analysis reveals that language grounding leads to more informative internal representations, better generalization to new partners, and improved capability for human-agent interaction. These findings demonstrate the effectiveness of integrating structured language into multi-agent learning and open avenues for more interpretable and capable multi-agent systems.
format Preprint
id arxiv_https___arxiv_org_abs_2506_05236
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Towards Language-Augmented Multi-Agent Deep Reinforcement Learning
Toquebiau, Maxime
Jun, Jae-Yun
Benamar, Faïz
Bredeche, Nicolas
Multiagent Systems
Most prior works on communication in multi-agent reinforcement learning have focused on emergent communication, which often results in inefficient and non-interpretable systems. Inspired by the role of language in natural intelligence, we investigate how grounding agents in a human-defined language can improve the learning and coordination of embodied agents. We propose a framework in which agents are trained not only to act but also to produce and interpret natural language descriptions of their observations. This language-augmented learning serves a dual role: enabling efficient and interpretable communication between agents, and guiding representation learning. We demonstrate that language-augmented agents outperform emergent communication baselines across various tasks. Our analysis reveals that language grounding leads to more informative internal representations, better generalization to new partners, and improved capability for human-agent interaction. These findings demonstrate the effectiveness of integrating structured language into multi-agent learning and open avenues for more interpretable and capable multi-agent systems.
title Towards Language-Augmented Multi-Agent Deep Reinforcement Learning
topic Multiagent Systems
url https://arxiv.org/abs/2506.05236