OpenCodePapers

atari-games-on-atari-2600-asteroids

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.
PaperCodeScoreModelNameReleaseDate
Generalized Data Distribution Iteration760005GDI-H32022-06-07
Generalized Data Distribution Iteration751970GDI-I32022-06-07
Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model✓ Link678558.64MuZero2019-11-19
Online and Offline Reinforcement Learning by Planning with a Learned Model✓ Link476412MuZero (Res2 Adam)2021-04-13
Recurrent Experience Replay in Distributed Reinforcement Learning✓ Link357867.7R2D22019-05-01
DNA: Proximal Policy Optimization with a Dual Network Architecture✓ Link165973DNA2022-06-20
Distributed Prioritized Experience Replay✓ Link155495.1Ape-X2018-03-02
Agent57: Outperforming the Atari Human Benchmark✓ Link150854.61Agent572020-03-30
IMPALA: Scalable Distributed Deep-RL with Importance Weighted Actor-Learner Architectures✓ Link108590.05IMPALA (deep)2018-02-05
Noisy Networks for Exploration✓ Link86700NoisyNet-Dueling2017-06-30
Mastering Atari with Discrete World Models✓ Link41526DreamerV22020-10-05
Evolving simple programs for playing Atari games✓ Link9412CGP2018-06-14
Asynchronous Methods for Deep Reinforcement Learning✓ Link5093.1A3C LSTM hs2016-02-04
The Arcade Learning Environment: An Evaluation Platform for General Agents✓ Link4660.6UCT2012-07-19
Fully Parameterized Quantile Function for Distributional Reinforcement Learning✓ Link4553.0FQF2019-11-05
Asynchronous Methods for Deep Reinforcement Learning✓ Link4474.5A3C FF hs2016-02-04
Distributional Reinforcement Learning with Quantile Regression✓ Link4226QR-DQN-12017-10-27
The Reactor: A fast and sample-efficient Actor-Critic agent for Reinforcement Learning3726.1Reactor 500M2017-04-15
Asynchronous Methods for Deep Reinforcement Learning✓ Link3009.4A3C FF (1 day) hs2016-02-04
Implicit Quantile Networks for Distributional Reinforcement Learning✓ Link2898IQN2018-06-14
Learning values across many orders of magnitude2869.3DDQN+Pop-Art noop2016-02-24
Dueling Network Architectures for Deep Reinforcement Learning✓ Link2837.7Duel noop2015-11-20
Prioritized Experience Replay✓ Link2654.3Prior noop2015-11-18
Policy Optimization With Penalized Point Probability Distance: An Alternative To Proximal Policy Optimization✓ Link2488.1POP3D2018-07-02
Self-Imitation Learning✓ Link2259.4A2C + SIL2018-06-14
Dueling Network Architectures for Deep Reinforcement Learning✓ Link2035.4Duel hs2015-11-20
Train a Real-world Local Path Planner in One Hour via Partially Decoupled Reinforcement Learning and Vectorized Diversity✓ Link1984.5ASL DDQN2023-05-07
Increasing the Action Gap: New Operators for Reinforcement Learning✓ Link1924.42Advantage Learning2015-12-15
Prioritized Experience Replay✓ Link1745.1Prior hs2015-11-18
Increasing the Action Gap: New Operators for Reinforcement Learning✓ Link1673.52Persistent AL2015-12-15
Human level control through deep reinforcement learning✓ Link1629.0Nature DQN2015-02-25
Evolution Strategies as a Scalable Alternative to Reinforcement Learning✓ Link1562.0ES FF (1 hour) noop2017-03-10
A Distributional Perspective on Reinforcement Learning✓ Link1516.0C51 noop2017-07-21
Deep Reinforcement Learning with Double Q-learning✓ Link1458.7DQN hs2015-09-22
Deep Reinforcement Learning with Double Q-learning✓ Link1364.5DQN noop2015-09-22
Deep Reinforcement Learning with Double Q-learning✓ Link1193.2DDQN (tuned) hs2015-09-22
Dueling Network Architectures for Deep Reinforcement Learning✓ Link1192.7Prior+Duel noop2015-11-20
Deep Exploration via Bootstrapped DQN✓ Link1032Bootstrapped DQN2016-02-15
Deep Reinforcement Learning with Double Q-learning✓ Link1021.9Prior+Duel hs2015-09-22
Massively Parallel Methods for Deep Reinforcement Learning✓ Link933.6Gorila2015-07-15
The Arcade Learning Environment: An Evaluation Platform for General Agents✓ Link907.3Best Learner2012-07-19
Dueling Network Architectures for Deep Reinforcement Learning✓ Link734.7DDQN (tuned) noop2015-11-20
[]()89.0SARSA
Check out how to contribute new results.