Atari Games On Atari 2600 Gopher

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
QR-DQN-1	113585	Distributional Reinforcement Learning with Quantile Regression
Prior hs	34858.8	Prioritized Experience Replay
A3C FF hs	10022.8	Asynchronous Methods for Deep Reinforcement Learning
UCT	20560	The Arcade Learning Environment: An Evaluation Platform for General Agents
Nature DQN	8520.0	Human level control through deep reinforcement learning	-
Gorila	4373.0	Massively Parallel Methods for Deep Reinforcement Learning
DQN noop	8777.4	Deep Reinforcement Learning with Double Q-learning
ES FF (1 hour) noop	582.0	Evolution Strategies as a Scalable Alternative to Reinforcement Learning
SARSA	2368.0	-	-
IMPALA (deep)	66782.30	IMPALA: Scalable Distributed Deep-RL with Importance Weighted Actor-Learner Architectures
Prior+Duel noop	104368.2	Dueling Network Architectures for Deep Reinforcement Learning
Duel noop	15718.4	Dueling Network Architectures for Deep Reinforcement Learning
POP3D	6207	Policy Optimization With Penalized Point Probability Distance: An Alternative To Proximal Policy Optimization
DDQN (tuned) noop	14840.8	Dueling Network Architectures for Deep Reinforcement Learning
GDI-H3	473560	Generalized Data Distribution Iteration	-
DDQN+Pop-Art noop	56218.2	Learning values across many orders of magnitude	-
CGP	1696	Evolving simple programs for playing Atari games
DDQN (tuned) hs	15253.0	Deep Reinforcement Learning with Double Q-learning
Persistent AL	10611.81	Increasing the Action Gap: New Operators for Reinforcement Learning
MuZero (Res2 Adam)	122882.5	Online and Offline Reinforcement Learning by Planning with a Learned Model

0 of 43 row(s) selected.