Timezone: »
Actor-critic methods, a type of model-free Reinforcement Learning, have been successfully applied to challenging tasks in continuous control, often achieving state-of-the art performance. However, wide-scale adoption of these methods in real-world domains is made difficult by their poor sample efficiency. We address this problem both theoretically and empirically. On the theoretical side, we identify two phenomena preventing efficient exploration in existing state-of-the-art algorithms such as Soft Actor Critic. First, combining a greedy actor update with a pessimistic estimate of the critic leads to the avoidance of actions that the agent does not know about, a phenomenon we call pessimistic underexploration. Second, current algorithms are directionally uninformed, sampling actions with equal probability in opposite directions from the current mean. This is wasteful, since we typically need actions taken along certain directions much more than others. To address both of these phenomena, we introduce a new algorithm, Optimistic Actor Critic, which approximates a lower and upper confidence bound on the state-action value function. This allows us to apply the principle of optimism in the face of uncertainty to perform directed exploration using the upper bound while still using the lower bound to avoid overestimation. We evaluate OAC in several challenging continuous control tasks, achieving state-of the art sample efficiency.
Author Information
Kamil Ciosek (Microsoft)
Quan Vuong (University of California San Diego)
Robert Loftin (Microsoft Research)
Katja Hofmann (Microsoft Research)
Dr. Katja Hofmann is a Principal Researcher at the [Game Intelligence](http://aka.ms/gameintelligence/) group at [Microsoft Research Cambridge, UK](https://www.microsoft.com/en-us/research/lab/microsoft-research-cambridge/). There, she leads a research team that focuses on reinforcement learning with applications in modern video games. She and her team strongly believe that modern video games will drive a transformation of how we interact with AI technology. One of the projects developed by her team is [Project Malmo](https://www.microsoft.com/en-us/research/project/project-malmo/), which uses the popular game Minecraft as an experimentation platform for developing intelligent technology. Katja's long-term goal is to develop AI systems that learn to collaborate with people, to empower their users and help solve complex real-world problems. Before joining Microsoft Research, Katja completed her PhD in Computer Science as part of the [ILPS](https://ilps.science.uva.nl/) group at the [University of Amsterdam](https://www.uva.nl/en). She worked with Maarten de Rijke and Shimon Whiteson on interactive machine learning algorithms for search engines.
Related Events (a corresponding poster, oral, or spotlight)
-
2019 Poster: Better Exploration with Optimistic Actor Critic »
Wed Dec 11th 01:30 -- 03:30 AM Room East Exhibition Hall B + C
More from the Same Authors
-
2020 Workshop: Competition Track Saturday »
Hugo Jair Escalante · Katja Hofmann -
2020 Workshop: Competition Track Friday »
Hugo Jair Escalante · Katja Hofmann -
2020 Poster: Multi-task Batch Reinforcement Learning with Metric Learning »
Jiachen Li · Quan Vuong · Shuang Liu · Minghua Liu · Kamil Ciosek · Henrik Christensen · Hao Su -
2019 Poster: Generalization in Reinforcement Learning with Selective Noise Injection and Information Bottleneck »
Maximilian Igl · Kamil Ciosek · Yingzhen Li · Sebastian Tschiatschek · Cheng Zhang · Sam Devlin · Katja Hofmann -
2019 Poster: Successor Uncertainties: Exploration and Uncertainty in Temporal Difference Learning »
David Janz · Jiri Hron · Przemysław Mazur · Katja Hofmann · José Miguel Hernández-Lobato · Sebastian Tschiatschek -
2019 Tutorial: Reinforcement Learning: Past, Present, and Future Perspectives »
Katja Hofmann -
2016 Demonstration: Project Malmo - Minecraft for AI Research »
Katja Hofmann · Matthew A Johnson · Fernando Diaz · Alekh Agarwal · Tim Hutton · David Bignell · Evelyne Viegas