Atari Games On Atari 2600 Ice Hockey
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 |
|---|---|
| deep-reinforcement-learning-with-double-q | 0.5 |
| impala-scalable-distributed-deep-rl-with | 3.48 |
| dna-proximal-policy-optimization-with-a-dual | 7.2 |
| human-level-control-through-deep | -1.6 |
| noisy-networks-for-exploration | 3 |
| deep-reinforcement-learning-with-double-q | -1.6 |
| mastering-atari-go-chess-and-shogi-by | 67.04 |
| evolving-simple-programs-for-playing-atari | 4 |
| implicit-quantile-networks-for-distributional | 0.2 |
| generalized-data-distribution-iteration | 44.94 |
| a-distributional-perspective-on-reinforcement | -3.5 |
| the-arcade-learning-environment-an-evaluation | -9.5 |
| mastering-atari-with-discrete-world-models-1 | 26 |
| agent57-outperforming-the-atari-human | 63.64 |
| self-imitation-learning | -2.4 |
| distributional-reinforcement-learning-with-1 | -1.7 |
| asynchronous-methods-for-deep-reinforcement | -1.7 |
| train-a-real-world-local-path-planner-in-one | -3.6 |
| prioritized-experience-replay | -0.2 |
| asynchronous-methods-for-deep-reinforcement | -4.7 |
| policy-optimization-with-penalized-point | -4.12 |
| gdi-rethinking-what-makes-reinforcement | 44.94 |
| dueling-network-architectures-for-deep | -2.7 |
| generalized-data-distribution-iteration | 481.9 |
| evolution-strategies-as-a-scalable | -4.1 |
| asynchronous-methods-for-deep-reinforcement | -2.8 |
| massively-parallel-methods-for-deep | -1.7 |
| learning-values-across-many-orders-of | -4.1 |
| recurrent-experience-replay-in-distributed | 79.3 |
| 模型 30 | -3.2 |
| dueling-network-architectures-for-deep | -0.4 |
| increasing-the-action-gap-new-operators-for | -1.24 |
| dueling-network-architectures-for-deep | 0.5 |
| deep-exploration-via-bootstrapped-dqn | -1.3 |
| deep-reinforcement-learning-with-double-q | -1.9 |
| the-arcade-learning-environment-an-evaluation | 39.4 |
| fully-parameterized-quantile-function-for | 17.3 |
| dueling-network-architectures-for-deep | -1.3 |
| online-and-offline-reinforcement-learning-by | 41.66 |
| deep-reinforcement-learning-with-double-q | -2.5 |
| increasing-the-action-gap-new-operators-for | -0.25 |
| distributed-prioritized-experience-replay | 33 |
| prioritized-experience-replay | 1.3 |