Reinforcement Learning on Atari 2600 (test) (20 Game Evaluation)
15,133AlienDelta Net
Evaluation results
| Method | Links | ||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Delta NetEnvironment Steps=200 M2021.06 | 15,133 | 432 | 7,525 | 19,907 | 333 | 321 | 28,594 | 78,601 | 86,168 | 54,336 | 15,064 | 4,180 | 21 | 521,839 | 7.7 | 12,790 | 86,132 | 11,734 | 208,563 | 10,523 | |
| Delta RNNEnvironment Steps=200 M2021.06 | 12,177 | 269 | 7,670 | 17,180 | 332 | 328 | 28,973 | 92,205 | 101,974 | 53,344 | 15,068 | 3,500 | 21 | 10,381 | 7.5 | 13,898 | 58,181 | 11,635 | 159,296 | 11,365 | |
| RDNEnvironment Steps=200 M2021.06 | 11,220 | 832 | 8,647 | 10,980 | 348 | 328 | 23,934 | 67,039 | 86,008 | 32,923 | 20,319 | 4,168 | 21 | 987,275 | 7.9 | 4,373 | 170,871 | 13,026 | 240,003 | 9,126 | |
| LSTMEnvironment Steps=200 M2021.06 | 6,184 | 349 | 7,977 | 24,873 | 1,150 | 327 | 18,024 | 69,750 | 124,914 | 25,106 | 24,447 | 5,431 | 21 | 11,513 | 7.6 | 12,643 | 137,657 | 13,194 | 251,157 | 11,619 | |
| LTEnvironment Steps=200 M2021.06 | 4,751 | 646 | 6,465 | 2,667 | 480 | 317 | 22,444 | 57,715 | 48,261 | 16,223 | 13,969 | 3,052 | 21 | 8,389 | 7.5 | 12,425 | 2,333 | 12,035 | 252,555 | 7,371 | |
| FeedforwardEnvironment Steps=200 M2021.06 | 3,816 | 433 | 6,407 | 60,527 | 343 | 331 | 21,873 | 74,904 | 61,350 | 56,459 | 12,292 | 2,499 | 21 | 8,655 | 7.8 | 667 | 53,455 | 11,564 | 185,632 | 11,960 |