Deployment-Efficient Reinforcement Learning via Model-Based Offline Optimization | HTML5

Most reinforcement learning (RL) algorithms assume online access to the environment, in which one may readily interleave updates to the policy with experience collection using that policy. However, in many real-world applications such as health, education, dialogue agents, and robotics, the cost or potential risk of deploying a new data-collection policy is high, to the point that it can become prohibitive to update the data-collection policy more than a few times during learning. With this view, we propose a novel concept of deployment efficiency, measuring the number of distinct data-collection policies that are used during policy learning. We observe that naïvely applying existing model-free offline RL algorithms recursively does not lead to a practical deployment-efficient and sample-efficient algorithm. We propose a novel model-based algorithm, Behavior-Regularized Model-ENsemble (BREMEN) that can effectively optimize a policy offline using 10-20 times fewer data than prior works.

Deployment-Efficient Reinforcement Learning via Model-Based Offline Optimization Tatsuya Matsushima Hiroki Furuta 1 1 footnotemark: 1 The University of Tokyo {matsushima, furuta}@weblab.t.u-tokyo.ac.jp Yutaka Matsuo The University of Tokyo matsuo@weblab.t.u-tokyo.ac.jp &Ofir Nachum Google Research ofirnachum@google.com &Shixiang Shane Gu Google Research shanegu@google.com Equal contribution. Abstract Most reinforcement learning (RL) algorithms assume online access to the environment, in which one may readily interleave updates to the policy with experience collection using that policy. However

related reading

• Exploration for the Efficient Deployment of Reinforcement Learning Agentsopenreview.net
• Part 2: Kinds of RL Algorithms - Spinning Up documentationspinningup.openai.com
• [2005.01643] Offline Reinforcement Learning: Tutorial, Review, and Perspectives on Open Problemsar5iv.labs.arxiv.org
• [2202.06450] Towards Deployment-Efficient Reinforcement Learning: Lower Bound and Optimalityar5iv.labs.arxiv.org
• A (Long) Peek into Reinforcement Learning | Lil'Loglilianweng.github.io
• On-Policy Distillation - Thinking Machines Labthinkingmachines.ai
• [1805.12114] Deep Reinforcement Learning in a Handful of Trials using Probabilistic Dynamics Modelsar5iv.labs.arxiv.org
• [2204.05618] When Should We Prefer Offline Reinforcement Learning Over Behavioral Cloning?ar5iv.labs.arxiv.org
• Key Papers in Deep RL - Spinning Up documentationspinningup.openai.com
• [2307.04354] Policy Finetuning in Reinforcement Learning via Design of Experiments using Offline Dataar5iv.labs.arxiv.org
• [2201.11861] The Challenges of Exploration for Offline Reinforcement Learningar5iv.labs.arxiv.org
• [1802.09081] Temporal Difference Models: Model-Free Deep RL for Model-Based Controlar5iv.labs.arxiv.org