Incentivizing Exploration In Reinforcement Learning With Deep Predictive Models

TL;DR

This paper introduces a scalable exploration method for deep reinforcement learning in high-dimensional Atari games, using learned system dynamics to generate exploration bonuses, leading to improved performance over prior strategies.

Contribution

The paper proposes a novel exploration bonus approach based on a neural network model of system dynamics, enhancing exploration efficiency in complex, high-dimensional environments.

Findings

Most consistent improvement across challenging Atari games

Development of an AUC-100 metric for exploration evaluation

Scalable exploration bonuses effective in raw pixel input domains

Abstract

Achieving efficient and scalable exploration in complex domains poses a major challenge in reinforcement learning. While Bayesian and PAC-MDP approaches to the exploration problem offer strong formal guarantees, they are often impractical in higher dimensions due to their reliance on enumerating the state-action space. Hence, exploration in complex domains is often performed with simple epsilon-greedy methods. In this paper, we consider the challenging Atari games domain, which requires processing raw pixel inputs and delayed rewards. We evaluate several more sophisticated exploration strategies, including Thompson sampling and Boltzman exploration, and propose a new exploration method based on assigning exploration bonuses from a concurrently learned model of the system dynamics. By parameterizing our learned model with a neural network, we are able to develop a scalable and efficient approach to exploration bonuses that can be applied to tasks with complex, high-dimensional state spaces. In the Atari domain, our method provides the most consistent improvement across a range of games that pose a major challenge for prior methods. In addition to raw game-scores, we also develop an AUC-100 metric for the Atari Learning domain to evaluate the impact of exploration on this benchmark.