Reinforcement Learning on MinAtar SpaceInvaders (test)
101.42Mean Episodic RewardEEDQN
Evaluation results
| Method | Links | ||
|---|---|---|---|
| EEDQNevaluation_window=last 100 training episodes, seeds=102025.06 | 101.42 | — | |
| MaxMinDQNevaluation_window=last 100 training episodes, seeds=102025.06 | 90.18 | 0.05 | |
| AvgDQNevaluation_window=last 100 training episodes, seeds=102025.06 | 68.55 | 0.01 | |
| ESDQNevaluation_window=last 100 training episodes, seeds=102025.06 | 68.21 | 0.01 | |
| DQNevaluation_window=last 100 training episodes, seeds=102025.06 | 33.96 | 0.01 | |
| DDQNevaluation_window=last 100 training episodes, seeds=102025.06 | 24.71 | 0.01 |