Atari Games On Atari 2600 Venture
评估指标
Score
评测结果
各个模型在此基准测试上的表现结果
比较表格
| 模型名称 | Score |
|---|---|
| train-a-real-world-local-path-planner-in-one | 291 |
| deep-reinforcement-learning-with-double-q | 136.0 |
| distributional-reinforcement-learning-with-1 | 43.9 |
| generalized-data-distribution-iteration | 2000 |
| the-arcade-learning-environment-an-evaluation | 66 |
| increasing-the-action-gap-new-operators-for | 198.69 |
| online-and-offline-reinforcement-learning-by | 1731.47 |
| distributed-prioritized-experience-replay | 1813 |
| human-level-control-through-deep | 380.0 |
| exploration-by-self-supervised-exploitation | 2138 |
| count-based-exploration-in-feature-space-for | 1169.2 |
| 模型 12 | 0.6 |
| count-based-exploration-with-neural-density | 82.2 |
| deep-exploration-via-bootstrapped-dqn | 212.5 |
| first-return-then-explore | 2281 |
| count-based-exploration-with-neural-density | 48.0 |
| massively-parallel-methods-for-deep | 523.4 |
| a-distributional-perspective-on-reinforcement | 1520.0 |
| deep-reinforcement-learning-with-double-q | 21.0 |
| generalized-data-distribution-iteration | 2035 |
| impala-scalable-distributed-deep-rl-with | 0.00 |
| evolution-strategies-as-a-scalable | 760.0 |
| count-based-exploration-with-the-successor | 1241.8 |
| dueling-network-architectures-for-deep | 48.0 |
| exploration-by-random-network-distillation | 1859 |
| recurrent-experience-replay-in-distributed | 1970.7 |
| count-based-exploration-in-feature-space-for | 0.0 |
| dna-proximal-policy-optimization-with-a-dual | 0 |
| asynchronous-methods-for-deep-reinforcement | 25.0 |
| learning-values-across-many-orders-of | 1172.0 |
| rudder-return-decomposition-for-delayed | 1350 |
| large-scale-study-of-curiosity-driven | 416 |
| prioritized-experience-replay | 94.0 |
| asynchronous-methods-for-deep-reinforcement | 19.0 |
| dueling-network-architectures-for-deep | 497.0 |
| deep-reinforcement-learning-with-double-q | 29.0 |
| mastering-atari-with-discrete-world-models-1 | 2 |
| implicit-quantile-networks-for-distributional | 1318 |
| unifying-count-based-exploration-and | 0.0 |
| dueling-network-architectures-for-deep | 98.0 |
| asynchronous-methods-for-deep-reinforcement | 23.0 |
| deep-reinforcement-learning-with-double-q | 163.0 |
| evolving-simple-programs-for-playing-atari | 0 |
| exploration-by-self-supervised-exploitation | 1787 |
| exploration-a-study-of-count-based | 445.0 |
| policy-optimization-with-penalized-point | 36.33 |
| agent57-outperforming-the-atari-human | 2623.71 |
| noisy-networks-for-exploration | 815 |
| self-imitation-learning | 0 |
| prioritized-experience-replay | 54.0 |
| dueling-network-architectures-for-deep | 200.0 |
| generalized-data-distribution-iteration | 2000 |
| mastering-atari-go-chess-and-shogi-by | 0.40 |
| incentivizing-exploration-in-reinforcement | 0.0 |
| exploration-by-self-supervised-exploitation | 2188 |