Reinforcement Learning on Atari 2600 Hero ALE (test)
30,826.4Maximum Agent ScoreAverage Human
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Average HumanPure-exploration regime=true2022.06 | 30,826.4 | — | |
| BYOL-ExplorePure-exploration regime=true, Episodes=10, Seeds=32022.06 | 22,078.89 | — | |
| RNDPure-exploration regime=true, Episodes=10, Seeds=32022.06 | 16,964.23 | — | |
| ICMPure-exploration regime=true, Episodes=10, Seeds=32022.06 | 2,567.09 | — | |
| A2C+SILSelf-Imitation Learning=true2018.06 | — | 33,069 | |
| A3C-CTSExploration=CTS2018.06 | — | 15,210 | |
| Reactor-PixelCNNExploration=PixelCNN2018.06 | — | 28,000 |