Atari Games On Atari 2600 Gravitar

評価指標

Score

評価結果

このベンチマークにおける各モデルのパフォーマンス結果

モデル名	Score	Paper Title	Repository
SARSA	429.0	-	-
CGP	2350	Evolving simple programs for playing Atari games
Nature DQN	306.7	Human level control through deep reinforcement learning
ES FF (1 hour) noop	805.0	Evolution Strategies as a Scalable Alternative to Reinforcement Learning
Agent57	19213.96	Agent57: Outperforming the Atari Human Benchmark
SND-STD	4643	Self-supervised network distillation: an effective approach to exploration in sparse reward environments
MuZero	6682.70	Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model
SND-V	2741	Self-supervised network distillation: an effective approach to exploration in sparse reward environments
SND-VIC	6712	Self-supervised network distillation: an effective approach to exploration in sparse reward environments
A2C + SIL	1874.2	Self-Imitation Learning
A3C LSTM hs	320.0	Asynchronous Methods for Deep Reinforcement Learning
GDI-I3	5905	GDI: Rethinking What Makes Reinforcement Learning Different From Supervised Learning	-
DQNMMCe	1078.3	Count-Based Exploration with the Successor Representation
DDQN+Pop-Art noop	483.5	Learning values across many orders of magnitude	-
Duel noop	588.0	Dueling Network Architectures for Deep Reinforcement Learning
MuZero (Res2 Adam)	8006.93	Online and Offline Reinforcement Learning by Planning with a Learned Model
DQN hs	298.0	Deep Reinforcement Learning with Double Q-learning
IMPALA (deep)	359.50	IMPALA: Scalable Distributed Deep-RL with Importance Weighted Actor-Learner Architectures
C51 noop	440.0	A Distributional Perspective on Reinforcement Learning
GDI-H3	5915	Generalized Data Distribution Iteration	-

0 of 53 row(s) selected.