Atari Games On Atari 2600 Montezumas Revenge
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
比較表
| モデル名 | Score |
|---|---|
| recurrent-experience-replay-in-distributed | 2061.3 |
| generalized-data-distribution-iteration | 2500 |
| the-arcade-learning-environment-an-evaluation | 10.7 |
| incentivizing-exploration-in-reinforcement | 142 |
| count-based-exploration-in-feature-space-for | 2745.4 |
| dna-proximal-policy-optimization-with-a-dual | 0 |
| count-based-exploration-with-the-successor | 1778.8 |
| asynchronous-methods-for-deep-reinforcement | 53 |
| impala-scalable-distributed-deep-rl-with | 0.00 |
| deep-reinforcement-learning-with-double-q | 24.0 |
| asynchronous-methods-for-deep-reinforcement | 67 |
| gdi-rethinking-what-makes-reinforcement | 3000 |
| contingency-aware-exploration-in | 6635 |
| increasing-the-action-gap-new-operators-for | 0.42 |
| policy-optimization-with-penalized-point | 0 |
| human-level-control-through-deep | 0 |
| unifying-count-based-exploration-and | 3459 |
| train-a-real-world-local-path-planner-in-one | 0 |
| massively-parallel-methods-for-deep | 84 |
| count-based-exploration-in-feature-space-for | 399.5 |
| exploration-by-self-supervised-exploitation | 7838 |
| count-based-exploration-with-neural-density | 3705.5 |
| prioritized-experience-replay | 51 |
| distributed-prioritized-experience-replay | 2500.0 |
| dueling-network-architectures-for-deep | 22.0 |
| agent57-outperforming-the-atari-human | 9352.01 |
| exploration-by-random-network-distillation | 8152 |
| online-and-offline-reinforcement-learning-by | 2500 |
| evolving-simple-programs-for-playing-atari | 0 |
| asynchronous-methods-for-deep-reinforcement | 41 |
| exploration-by-self-supervised-exploitation | 7212 |
| exploration-a-study-of-count-based | 75 |
| exploration-by-self-supervised-exploitation | 21565 |
| count-based-exploration-with-the-successor | 1778.6 |
| self-imitation-learning | 1100 |
| large-scale-study-of-curiosity-driven | 2504.6 |
| first-return-then-explore | 43791 |
| unifying-count-based-exploration-and | 273.7 |
| deep-exploration-via-bootstrapped-dqn | 100 |
| モデル 40 | 259 |
| go-explore-a-new-approach-for-hard | 43763 |
| generalized-data-distribution-iteration | 3000 |
| deep-reinforcement-learning-with-double-q | 47.0 |
| mastering-atari-with-discrete-world-models-1 | 81 |
| noisy-networks-for-exploration | 57 |
| mastering-atari-go-chess-and-shogi-by | 0.00 |
| implicit-quantile-networks-for-distributional | 0 |
| distributional-reinforcement-learning-with-1 | 0 |
| increasing-the-action-gap-new-operators-for | 1.72 |
| deep-reinforcement-learning-with-double-q | 42.0 |