Reward-Free Exploration for Reinforcement Learning

by   Chi Jin, et al.

Exploration is widely regarded as one of the most challenging aspects of reinforcement learning (RL), with many naive approaches succumbing to exponential sample complexity. To isolate the challenges of exploration, we propose a new "reward-free RL" framework. In the exploration phase, the agent first collects trajectories from an MDP M without a pre-specified reward function. After exploration, it is tasked with computing near-optimal policies under for M for a collection of given reward functions. This framework is particularly suitable when there are many reward functions of interest, or when the reward function is shaped by an external agent to elicit desired behavior. We give an efficient algorithm that conducts Õ(S^2Apoly(H)/ϵ^2) episodes of exploration and returns ϵ-suboptimal policies for an arbitrary number of reward functions. We achieve this by finding exploratory policies that visit each "significant" state with probability proportional to its maximum visitation probability under any possible policy. Moreover, our planning procedure can be instantiated by any black-box approximate planner, such as value iteration or natural policy gradient. We also give a nearly-matching Ω(S^2AH^2/ϵ^2) lower bound, demonstrating the near-optimality of our algorithm in this setting.


page 1

page 2

page 3

page 4


Task-agnostic Exploration in Reinforcement Learning

Efficient exploration is one of the main challenges in reinforcement lea...

On Reward-Free Reinforcement Learning with Linear Function Approximation

Reward-free reinforcement learning (RL) is a framework which is suitable...

Exploration by Maximizing Rényi Entropy for Zero-Shot Meta RL

Exploring the transition dynamics is essential to the success of reinfor...

Nearly Minimax Optimal Reward-free Reinforcement Learning

We study the reward-free reinforcement learning framework, which is part...

On Reward-Free RL with Kernel and Neural Function Approximations: Single-Agent MDP and Markov Game

To achieve sample efficiency in reinforcement learning (RL), it necessit...

Safe Exploration Incurs Nearly No Additional Sample Complexity for Reward-free RL

While the primary goal of the exploration phase in reward-free reinforce...

Adaptive Reward-Free Exploration

Reward-free exploration is a reinforcement learning setting recently stu...