Conference item
Deep residual reinforcement learning
- Abstract:
-
We revisit residual algorithms in both model-free and model-based reinforcement learning settings. We propose the bidirectional target network technique to stabilize residual algorithms, yielding a residual version of DDPG that significantly outperforms vanilla DDPG in the DeepMind Control Suite benchmark. Moreover, we find the residual algorithm an effective approach to the distribution mismatch problem in model-based planning. Compared with the existing TD(k) method, our residual-based method makes weaker assumptions about the model and yields a greater performance boost.
- Publication status:
- Published
- Peer review status:
- Reviewed (other)
Actions
Authors
- Publisher:
- International Foundation for Autonomous Agents and Multiagent Systems
- Pages:
- 1611-1619
- Publication date:
- 2020-05-01
- Acceptance date:
- 2020-04-04
- Event title:
- Nineteenth International Conference on Autonomous Agents and Multi-Agent Systems
- Event location:
- Auckland, New Zealand
- Event website:
- https://aamas2020.conference.auckland.ac.nz/
- Event start date:
- 2020-05-09
- Event end date:
- 2020-05-13
- ISSN:
-
2523-5699
- ISBN:
- 978-1-4503-7518-4
- Language:
-
English
- Keywords:
- Pubs id:
-
1098604
- Local pid:
-
pubs:1098604
- Deposit date:
-
2020-04-04
Terms of use
- Copyright holder:
- International Foundation for Autonomous Agents and Multiagent Systems (IFAAMAS)
- Copyright date:
- 2020
- Rights statement:
- Copyright © 2020 by International Foundation for Autonomous Agents and Multiagent Systems (IFAAMAS).
- Notes:
-
This is the accepted manuscript version of the article. The final version is available from AAMS at http://www.ifaamas.org/Proceedings/aamas2020/pdfs/p1611.pdf
If you are the owner of this record, you can report an update to it here: Report update to this record