Learning To Reinforcement Learn

Abstract

In recent years deep reinforcement learning (RL) systems have attained superhuman performance in a number of challenging task domains. However, a major limitation of such applications is their demand for massive amounts of training data. A critical present objective is thus to develop deep RL methods that can adapt rapidly to new tasks. In the present work we introduce a novel approach to this challenge, which we refer to as deep meta-reinforcement learning. Previous work has shown that recurrent networks can support meta-learning in a fully supervised context. We extend this approach to the RL setting. What emerges is a system that is trained using one RL algorithm, but whose recurrent dynamics implement a second, quite separate RL procedure. This second, learned RL algorithm can differ from the original one in arbitrary ways. Importantly, because it is learned, it is configured to exploit structure in the training domain. We unpack these points in a series of seven proof-of-concept e

Stats

• citations0
• S2 citations—
• github stars0
• HF likes0
• heat score0.00
• arxiv keywang2016learning

Related papers

• A Tutorial On Meta-reinforcement Learning (2023)10.85
• Deep Online Learning Via Meta-learning: Continual Adaptation For Model-based RL (2018)0.00
• RL\(^3\): Boosting Meta Reinforcement Learning Via RL Inside RL\(^2\) (2023)0.00
• Context Meta-reinforcement Learning Via Neuromodulation (2021)6.34
• Meta-gradient Reinforcement Learning With An Objective Discovered Online (2020)0.00
• Improving Generalization In Meta Reinforcement Learning Using Learned Objectives (2019)0.00
• Unsupervised Meta-learning For Reinforcement Learning (2018)0.00
• Guided Meta-policy Search (2019)0.00