Munchausen Reinforcement Learning

TL;DR

This paper introduces a simple modification to reinforcement learning algorithms by adding the scaled log-policy to the reward, achieving state-of-the-art results on Atari games without complex techniques.

Contribution

The paper proposes a novel, straightforward approach to bootstrap RL using the current policy, improving performance and theoretical understanding with minimal changes to existing algorithms.

Findings

Competitive performance with distributional methods on Atari

Outperforms Rainbow with minimal modifications

Provides theoretical insights into implicit regularization and action-gap

Abstract

Bootstrapping is a core mechanism in Reinforcement Learning (RL). Most algorithms, based on temporal differences, replace the true value of a transiting state by their current estimate of this value. Yet, another estimate could be leveraged to bootstrap RL: the current policy. Our core contribution stands in a very simple idea: adding the scaled log-policy to the immediate reward. We show that slightly modifying Deep Q-Network (DQN) in that way provides an agent that is competitive with distributional methods on Atari games, without making use of distributional RL, n-step returns or prioritized replay. To demonstrate the versatility of this idea, we also use it together with an Implicit Quantile Network (IQN). The resulting agent outperforms Rainbow on Atari, installing a new State of the Art with very little modifications to the original algorithm. To add to this empirical study, we provide strong theoretical insights on what happens under the hood -- implicit Kullback-Leibler regularization and increase of the action-gap.