Learning Adaptive Exploration Strategies in Dynamic Environments Through Informed Policy Regularization

Abstract

We study the problem of learning exploration-exploitation strategies that effectively adapt to dynamic environments, where the task may change over time. While RNN-based policies could in principle represent such strategies, in practice their training time is prohibitive and the learning process often converges to poor solutions. In this paper, we consider the case where the agent has access to a description of the task (e.g., a task id or task parameters) at training time, but not at test time. We propose a novel algorithm that regularizes the training of an RNN-based policy using informed policies trained to maximize the reward in each task. This dramatically reduces the sample complexity of training RNN-based policies, without losing their representational power. As a result, our method learns exploration strategies that efficiently balance between gathering information about the unknown and changing task and maximizing the reward over time. We test the performance of our algorithm in a variety of environments where tasks may vary within each episode.

18 pages

References in corpus (8)

• Playing Atari with Deep Reinforcement Learning
• Asynchronous Methods for Deep Reinforcement Learning
• Efficient Off-Policy Meta-Reinforcement Learning via Probabilistic Context Variables
• On Upper-Confidence Bound Policies for Non-Stationary Bandit Problems
• Distral: Robust Multitask Reinforcement Learning
• Deep Exploration via Randomized Value Functions
• Rapid Learning or Feature Reuse? Towards Understanding the Effectiveness of MAML
• Meta reinforcement learning as task inference

Cited by in corpus (4)

• Offline Meta-Reinforcement Learning with Online Self-Supervision
• Decoupling Exploration and Exploitation for Meta-Reinforcement Learning without Sacrifices
• Block Contextual MDPs for Continual Learning
• Exploration in Approximate Hyper-State Space for Meta Reinforcement Learning