Reinforcement Learning on Atari 7-game suite ALE (test)
4.815Relative ScoreRainbow+AE
Evaluation results
| Method | Links | |
|---|---|---|
| Rainbow+AELearning Paradigm=Self-supervised2022.06 | 4.815 | |
| Eff.-RainbowLearning Paradigm=No SSL2022.06 | 2.065 | |
| BYOLLearning Paradigm=No SSL2022.06 | 1.877 | |
| Extract-RewardLearning Paradigm=Self-supervised2022.06 | 1.692 | |
| Predict-RewardLearning Paradigm=Self-supervised2022.06 | 0.135 | |
| DrQ*Learning Paradigm=No SSL, Image Augmentation=different from original paper2022.06 | -0.29 | |
| ELo-RainbowLearning Paradigm=Self-supervised2022.06 | -0.369 | |
| Predict-FR-BalancedLearning Paradigm=Self-supervised2022.06 | -0.508 | |
| CURLLearning Paradigm=No SSL2022.06 | -0.516 | |
| Predict-FutureLearning Paradigm=Self-supervised2022.06 | -1.796 | |
| Rainbow-AugLearning Paradigm=No SSL2022.06 | -2.222 | |
| Extract-ActionLearning Paradigm=Self-supervised2022.06 | -2.237 | |
| RotationCLSLearning Paradigm=Self-supervised2022.06 | -2.647 |