Abstract
We explore the benefits of augmenting state-of-the-art model-free deep reinforcement learning with simple object representations. Following the Frostbite challenge posited by Lake et al. (2017), we identify object representations as a critical cognitive capacity lacking from current reinforcement learning agents. We discover that providing the Rainbow model (Hessel et al., 2018) with simple, feature-engineered object representations substantially boosts its performance on the Frostbite game from Atari 2600. We then analyze the relative contributions of the representations of different types of objects, identify environment states where these representations are most impactful, and examine how these representations aid in generalizing to novel situations.
Original language | English (US) |
---|---|
Pages | 2023-2029 |
Number of pages | 7 |
State | Published - 2020 |
Event | 42nd Annual Meeting of the Cognitive Science Society: Developing a Mind: Learning in Humans, Animals, and Machines, CogSci 2020 - Virtual, Online Duration: Jul 29 2020 → Aug 1 2020 |
Conference
Conference | 42nd Annual Meeting of the Cognitive Science Society: Developing a Mind: Learning in Humans, Animals, and Machines, CogSci 2020 |
---|---|
City | Virtual, Online |
Period | 7/29/20 → 8/1/20 |
Keywords
- deep reinforcement learning
- DQN
- model-free reinforcement learning
- object representations
ASJC Scopus subject areas
- Artificial Intelligence
- Computer Science Applications
- Human-Computer Interaction
- Cognitive Neuroscience