A Survey of Exploration Methods in Reinforcement Learning

Abstract

Exploration is an essential component of reinforcement learning algorithms, where agents need to learn how to predict and control unknown and often stochastic environments. Reinforcement learning agents depend crucially on exploration to obtain informative data for the learning process as the lack of enough information could hinder effective learning. In this article, we provide a survey of modern exploration methods in (Sequential) reinforcement learning, as well as a taxonomy of exploration methods.

References in corpus (15)

• RL${}^2$: Fast Reinforcement Learning via Slow Reinforcement Learning
• Massively Parallel Methods for Deep Reinforcement Learning
• Learning to reinforcement learn
• PEGASUS: A Policy Search Method for Large MDPs and POMDPs
• Model-Based Bayesian Exploration
• Efficient Off-Policy Meta-Reinforcement Learning via Probabilistic Context Variables
• Path Integral Policy Improvement with Covariance Matrix Adaptation
• A Bayesian Sampling Approach to Exploration in Reinforcement Learning
• REGAL: A Regularization based Algorithm for Reinforcement Learning in Weakly Communicating MDPs
• A unified view of entropy-regularized Markov decision processes
• Provably Efficient Maximum Entropy Exploration
• Online Regret Bounds for Undiscounted Continuous Reinforcement Learning
• Better Exploration with Optimistic Actor-Critic
• Optimistic Exploration even with a Pessimistic Initialisation
• Discovering Options for Exploration by Minimizing Cover Time