About this document
MARLIN: Efficient Multi-Agent RL Training by Nguyễn Vân Anh is a document available to read on EtoBox.
The document introduces MARLIN, a method that enhances multi-agent reinforcement learning (MARL) by incorporating language-based inter-robot negotiation to reduce training time and improve deployment efficiency. By utilizing large language models for negotiation, robots can generate plans that guide their actions, allowing them to achieve peak performance with fewer training episodes compared to traditional MARL approaches. The effectiveness of this hybrid method is evaluated against existing MARL technique
- Author
- Nguyễn Vân Anh
- Language
- EN