Adversarially Guided Actor-Critic

Abstract

Despite definite success in deep reinforcement learning problems, actor-critic algorithms are still confronted with sample inefficiency in complex environments, particularly in tasks where efficient exploration is a bottleneck. These methods consider a policy (the actor) and a value function (the critic) whose respective losses are built using different motivations and approaches. This paper introduces a third protagonist: the adversary. While the adversary mimics the actor by minimizing the KL-divergence between their respective action distributions, the actor, in addition to learning to solve the task, tries to differentiate itself from the adversary predictions. This novel objective stimulates the actor to follow strategies that could not have been correctly predicted from previous trajectories, making its behavior innovative in tasks where the reward is extremely rare. Our experimental analysis shows that the resulting Adversarially Guided Actor-Critic (AGAC) algorithm leads to more exhaustive exploration. Notably, AGAC outperforms current state-of-the-art methods on a set of various hard-exploration and procedurally-generated tasks.

Conformal Bootstrap with Reinforcement Learning

“…Such algorithms can be split into two main sets and can be distinguished by whether the actions (defined by numbers) taken by the agent are discrete or continuous. Algorithms such as Deep Q-Learning [31] or Actor-Critic methods [32] use a discrete action space (convenient when one can take only a finite amount of actions), while algorithms such as the soft Actor-Critic method [22] and the Deep Deterministic Policy Gradient method [33] were developed for when the actions can take any real value.…”

Explore and Control with Adversarial Surprise

“…Emergence in multi-agent setting: Multi-agent competition can provide a mechanism for driving RL agents to automatically learn increasingly complex behavior [27]. As each agent adapts, it makes the learning problem for the other agent increasingly difficult, leading to the emergence of an automatic curriculum of challenging learning tasks [2,10,47,14]. For example, Schmidhuber [38] proposed having two classifiers compete by repeatedly selecting examples which they can classify but which the other cannot.…”