Model-Based Opponent Modeling

Abstract

When one agent interacts with a multi-agent environment, it is challenging to deal with various opponents unseen before. Modeling the behaviors, goals, or beliefs of opponents could help the agent adjust its policy to adapt to different opponents. In addition, it is also important to consider opponents who are learning simultaneously or capable of reasoning. However, existing work usually tackles only one of the aforementioned types of opponent. In this paper, we propose model-based opponent modeling (MBOM), which employs the environment model to adapt to all kinds of opponent. MBOM simulates the recursive reasoning process in the environment model and imagines a set of improving opponent policies. To effectively and accurately represent the opponent policy, MBOM further mixes the imagined opponent policies according to the similarity with the real behaviors of opponents. Empirically, we show that MBOM achieves more effective adaptation than existing methods in competitive and cooperative environments, respectively with different types of opponent, i.e., fixed policy, naïve learner, and reasoning learner.Preprint. Under review.

Consistent Epistemic Planning for Multiagent Deep Reinforcement Learning

“…Grover et al use imitation learning and contrastive learning to predict the next actions of an opponent [17]. Yu et al utilize a model-based approach with neural networks for predicting the next state of the opponents [18]. We differ from prior approaches by 1) only utilizing sparse observations (i.e., we do not have access to the true states of opponents unless detected) and 2) we justify the value of the designed filter by helping train a MARL to complete S&T tasks which could not be done without the filter as shown in §V-B.…”