MARLIN introduces a novel approach where robots use language-based negotiation during reinforcement learning to improve task outcomes and safety.

• Enables robots to communicate decisions and reasoning before taking actions
• Reduces training episodes needed by up to 50% through negotiation-based optimization
• Significantly improves safety and task completion rates in complex multi-robot environments
• Creates more interpretable robot decision-making for human operators

This engineering breakthrough makes multi-robot systems more practical for real-world deployment by reducing training time and enhancing reliability in critical applications like disaster response and industrial automation.

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