Evaluating Generalization in Multiagent Systems using Agent-Interaction Graphs

Learning from interactions between agents is a key component for inference in multiagent systems. Depending on the downstream task, there could be multiple criteria for evaluating the generalization performance of learning. In this work, we propose a novel framework for evaluating generalization in multiagent systems based on agent-interaction graphs. An agent-interaction graph models agents as nodes and interactions as hyper-edges between participating agents. Using this abstract data structure, we define three notions of generalization for principled evaluation of learning in multiagent systems.

Authors

• Aditya Grover University of California, Los Angeles
• Maruan Al-Shedivat King Abdullah University of Science and Technology
• Jayesh K. Gupta
• Yuri Burda OpenAI
• Harrison Edwards OpenAI

Keywords

• Generalization
• multiagent systems
• agent-interaction graphs