MARLIN: Multi-Agent Reinforcement Learning Guided by Language-Based Inter-Robot Negotiation
MARLIN: Multi-Agent Reinforcement Learning Guided by Language-Based Inter-Robot Negotiation
This research introduces MARLIN, a hybrid framework that integrates large language models for inter-robot negotiation to guide multi-agent reinforcement learning. The system dynamically switches between language-based planning and RL policy learning, achieving higher performance during early training stages while maintaining final performance comparable to standard methods. Evaluated on both simulated and physical robots with local and remote language models, the code is publicly available.
The paper proposes MARLIN, a hybrid framework combining large language models for high-level negotiation with multi-agent reinforcement learning to improve early-stage training performance in both simulated and physical robot settings.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
- code: Available Artifact link Open source S4
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
ing training, enabling 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 with limited impact on the final performance. The code is available at https://github.com/SooratiLab/MARLIN.
Open source S4
Source：arXiv Robotics — research abstracts · arxiv.org