Action Guidance with MCTS for Deep Reinforcement Learning
This addresses sample inefficiency for reinforcement learning practitioners in domains with sparse, delayed, and deceptive rewards, but it is incremental as it builds on existing methods.
The paper tackles sample inefficiency in deep reinforcement learning by integrating action guidance from a non-expert simulated demonstrator, such as Monte Carlo tree search, into asynchronous distributed methods, resulting in faster learning and better policies on a two-player mini version of the Pommerman game.
Deep reinforcement learning has achieved great successes in recent years, however, one main challenge is the sample inefficiency. In this paper, we focus on how to use action guidance by means of a non-expert demonstrator to improve sample efficiency in a domain with sparse, delayed, and possibly deceptive rewards: the recently-proposed multi-agent benchmark of Pommerman. We propose a new framework where even a non-expert simulated demonstrator, e.g., planning algorithms such as Monte Carlo tree search with a small number rollouts, can be integrated within asynchronous distributed deep reinforcement learning methods. Compared to a vanilla deep RL algorithm, our proposed methods both learn faster and converge to better policies on a two-player mini version of the Pommerman game.