Reinforcement Learning on Atari 2600 25 games
518.2Mean Human Normalized ScoreMBEC++
Evaluation results
| Method | Links | ||
|---|---|---|---|
| MBEC++Training frames=10 million2021.11 | 518.2 | 133.4 | |
| ERLAMTraining frames=10 million2021.11 | 515.4 | 103.5 | |
| EMDQNTraining frames=40 million2021.11 | 250.6 | 95.5 | |
| EVATraining frames=10 million2021.11 | 172.2 | 39.2 | |
| NECTraining frames=10 million2021.11 | 106.1 | 53.3 | |
| Nature DQNTraining frames=10 million2021.11 | 83.6 | 16 | |
| MFECTraining frames=10 million2021.11 | 77.7 | 40.9 |