InfoRL: Interpretable Reinforcement Learning using Information Maximization

Recent advances in reinforcement learning have proved that given an\nenvironment we can learn to perform a task in that environment if we have\naccess to some form of a reward function (dense, sparse or derived from IRL).\nBut most of the algorithms focus on learning a single best policy to perform a\ngiven set of tasks. In this paper, we focus on an algorithm that learns to not\njust perform a task but different ways to perform the same task. As we know\nwhen the environment is complex enough there always exists multiple ways to\nperform a task. We show that using the concept of information maximization it\nis possible to learn latent codes for discovering multiple ways to perform any\ngiven task in an environment.\n