Coordination Graph

Coordination graphs represent relationships between agents or data points, aiming to model collaboration or identify outliers within complex systems. Current research focuses on developing efficient algorithms for learning these graphs, particularly in multi-agent reinforcement learning, often employing graph neural networks and incorporating temporal or group-level information to improve coordination and scalability. These advancements are significant for improving the performance of multi-agent systems in various applications, such as resource allocation and robotics, and for enhancing outlier detection methods in data analysis.

Papers