{"type":"ScholarlyArticle","title":"Deep Residual Reinforcement Learning (Extended Abstract)","authors":[{"name":"Shangtong Zhang","slug":"shangtong-zhang","given":"Shangtong","family":"Zhang","affiliation":""},{"name":"Wendelin Boehmer","slug":"wendelin-boehmer","given":"Wendelin","family":"Boehmer","affiliation":""},{"name":"Shimon Whiteson","slug":"shimon-whiteson","given":"Shimon","family":"Whiteson","affiliation":""}],"abstract":"We revisit residual algorithms in both model-free and model-based reinforcement learning settings. \r\nWe propose the bidirectional target network technique to stabilize residual algorithms, yielding a residual version of DDPG that significantly outperforms vanilla DDPG in commonly used benchmarks. \r\nMoreover, we find the residual algorithm an effective approach to the distribution mismatch problem in model-based planning. \r\nCompared with the existing TD(k) method, our residual-based method makes weaker assumptions about the model and yields a greater performance boost.","doi":"10.24963/ijcai.2021/668","url":"https://ijcai.org/proceedings/2021/668","pdf":"https://ijcai.org/proceedings/2021/0668.pdf","markdown":"https://ijcai.org/proceedings/2021/668.md","bibtex":"https://ijcai.org/proceedings/2021/bibtex/668","volume":{"slug":"2021","title":"Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence","short_title":"IJCAI 2021","url":"https://ijcai.org/proceedings/2021/","year":2021,"location":"Montreal","dates":"19-27 August 2021","editors":"Prof. Zhi-Hua Zhou","isbn":"978-0-9992411-9-6","publisher":"International Joint Conferences on Artificial Intelligence Organization","doi_prefix":"10.24963/ijcai.2021"},"section":"Sister Conferences Best Papers","track":"Sister Conferences Best Papers","number":668,"pages":{"first":4869,"last":4873},"keywords":["Machine Learning: Reinforcement Learning"],"videos":[],"references":[],"license_note":"Copyright International Joint Conferences on Artificial Intelligence."}