Reinforcement Learning on Atari 2600 random start condition
6,072Beam Rider ScoreMAC
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| MACCondition=random start condition, Training Frames=50 million, Trials=52017.09 | 6,072 | 372.7 | 10.6 | 243.4 | 1,703.4 | 1,173.1 | |
| A2CCondition=random start condition, Training Frames=50 million, Trials=52017.09 | 5,846 | 370.9 | 18 | 1,651.5 | 1,702.5 | 1,201.2 | |
| TRPOCondition=random start condition, Trials=12017.09 | 1,425.2 | 10.8 | 20.9 | 1,973.5 | 1,908.6 | 568.4 | |
| ESCondition=random start condition, Trials=302017.09 | 744 | 9.5 | 21 | 147.5 | 1,390 | 678.5 | |
| RandomCondition=random start condition2017.09 | 363.9 | 1.7 | -20.7 | 183 | 68.4 | 148 |