MARLIN: Multi-Agent Reinforcement Learning Guided by Language-Based Inter-Robot Negotiation

TL;DR

MARLIN integrates language-based negotiation with reinforcement learning to improve multi-robot training safety and efficiency, enabling better early-stage performance and safer exploration.

Contribution

The paper presents a hybrid framework combining language models with reinforcement learning for multi-robot systems, enhancing early training safety and performance.

Findings

Hybrid approach outperforms standard RL in early training stages.

Language-based negotiation improves safety during initial learning.

System achieves higher early performance without sacrificing final results.

Abstract

Multi-agent reinforcement learning is a key method for training multi-robot systems. Through rewarding or punishing robots over a series of episodes according to their performance, they can be trained and then deployed in the real world. However, poorly trained policies can lead to unsafe behaviour during early training stages. We introduce Multi-Agent Reinforcement Learning guided by language-based Inter-robot Negotiation (MARLIN), a hybrid framework in which large language models provide high-level planning before the reinforcement learning policy has learned effective behaviours. Robots use language models to negotiate actions and generate plans that guide policy learning. The system dynamically switches between reinforcement learning and language-model-based negotiation during training, enabling safer and more effective exploration. MARLIN is evaluated using both simulated and physical robots with local and remote language models. Results show that, compared to standard multi-agent reinforcement learning, the hybrid approach achieves higher performance in early training without reducing final performance. The code is available at https://github.com/SooratiLab/MARLIN.