OpenCodePapers

atari-games-on-atari-57

Video GamesAtari Games
Dataset Link
Results over time
Click legend items to toggle metrics. Hover points for model names.
Leaderboard
Click a sortable column header to sort.
PaperCodeMean Human Normalized ScoreHuman World Record BreakthroughModelNameReleaseDate
Learnable Behavior Control: Breaking Atari Human World Records via Sample-Efficient Behavior Selection10077.52%24LBC2023-05-09
GDI: Rethinking What Makes Reinforcement Learning Different From Supervised Learning9620.98%22GDI-H3(200M frames)2021-06-11
Generalized Data Distribution Iteration9620.33%22GDI-H32022-06-07
Generalized Data Distribution Iteration7810.1%17GDI-I32022-06-07
Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model✓ Link4996.20%19MuZero2019-11-19
Recurrent Experience Replay in Distributed Reinforcement Learning✓ Link3374.31%15R2D22019-05-01
Off-Policy Actor-Critic with Shared Experience Replay1741.36%7LASER2019-09-25
IMPALA: Scalable Distributed Deep-RL with Importance Weighted Actor-Learner Architectures✓ Link957.34%3IMPALA, deep2018-02-05
Rainbow: Combining Improvements in Deep Reinforcement Learning✓ Link873.97%4Rainbow DQN2017-10-06
Munchausen Reinforcement Learning✓ Link504%M-IQN2020-07-28
GDI: Rethinking What Makes Reinforcement Learning Different From Supervised LearningGDI-H32021-06-11
Check out how to contribute new results.