Provably Efficient Exploration for Reinforcement Learning Using Unsupervised Learning

Part of Advances in Neural Information Processing Systems 33 pre-proceedings (NeurIPS 2020)

Bibtex »Paper »Supplemental »

Bibtek download is not availble in the pre-proceeding


Authors

Fei Feng, Ruosong Wang, Wotao Yin, Simon S. Du, Lin Yang

Abstract

<p>Motivated by the prevailing paradigm of using unsupervised learning for efficient exploration in reinforcement learning (RL) problems [tang2017exploration,bellemare2016unifying], we investigate when this paradigm is provably efficient. We study episodic Markov decision processes with rich observations generated from a small number of latent states. We present a general algorithmic framework that is built upon two components: an unsupervised learning algorithm and a no-regret tabular RL algorithm. Theoretically, we prove that as long as the unsupervised learning algorithm enjoys a polynomial sample complexity guarantee, we can find a near-optimal policy with sample complexity polynomial in the number of latent states, which is significantly smaller than the number of observations. Empirically, we instantiate our framework on a class of hard exploration problems to demonstrate the practicality of our theory.</p>