Reinforcement Learning on Atari 2600 (test) (Full Game Set)
4,704Alien ScoreDelta Net
Evaluation Results
| Method | Links | ||||||||||||||||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Delta NetEnvironmental steps=50 M, Number of parameters=1.5 M2021.06 | 4,704 | 339 | 5,710 | — | — | — | 321 | 6,780 | — | — | — | — | — | — | — | 31,359 | — | — | — | — | — | 8,707 | — | — | 3,338 | — | — | 8,095 | — | 2,116 | — | — | 21 | — | 6,248 | 9,185 | — | 7.2 | 8,853 | 25,769 | 11,599 | — | — | — | 120,806 | — | — | — | 4,265 | 331 | |
| Delta RNNEnvironmental steps=50 M, Number of parameters=1.6 M2021.06 | 3,420 | 216 | 7,503 | — | — | — | 324 | 7,040 | — | — | — | — | — | — | — | 36,807 | — | — | — | — | — | 25,445 | — | — | 2,929.3 | — | — | 7,388 | — | 2,287 | — | — | 21 | — | 6,626 | 13,429 | — | 7 | 5,123 | 2,847 | 9,523 | — | — | — | 148,759 | — | — | — | 3,903 | 305 | |
| LTEnvironmental steps=50 M, Number of parameters=1.5 M2021.06 | 2,135 | 320 | 2,764 | — | — | — | 309 | 933 | — | — | — | — | — | — | — | 8,939 | — | — | — | — | — | 8,197 | — | — | 2,425 | — | — | 3,722 | — | 1,647 | — | — | 21 | — | 4,693 | 6,695 | — | 4.8 | 1,812 | 2,345 | 8,915 | — | — | — | 57,435 | — | — | — | 2,719 | 323 | |
| FeedforwardEnvironmental steps=50 M, Number of parameters=1.1 M2021.06 | 1,985 | 208 | 4,658 | — | — | — | 323 | 12,267 | — | — | — | — | — | — | — | 36,255 | — | — | — | — | — | 10,356 | — | — | 2,942 | — | — | 5,449 | — | 1,737 | — | — | 21 | — | 4,967 | 9,932 | — | 7.1 | 469 | 48,150 | 9,397 | — | — | — | 185,632 | — | — | — | 4,863 | 326 | |
| RDNEnvironmental steps=50 M, Number of parameters=1.5 M2021.06 | 1,754 | 368 | 8,088 | — | — | — | 316 | 7,373 | — | — | — | — | — | — | — | 41,726 | — | — | — | — | — | 19,775 | — | — | 2,979 | — | — | 9,201 | — | 2,584 | — | — | 21 | — | 5,897 | 18,156 | — | 7.5 | 686 | 27,213 | 9,737 | — | — | — | 126,140 | — | — | — | 5,285 | 336 | |
| POP3DFrames=40M, Trials=32018.07 | 1,147.29 | 299.55 | 2,139.15 | 2,004.43 | 1,652.48 | 488,134.03 | 662.26 | 11,131.44 | 1,965.27 | 37.97 | 83.12 | 143.6 | 3,056.81 | 3,269.47 | 97,257.52 | 7,611.27 | — | 107.84 | -21 | 17.76 | 276.47 | 1,556.29 | 413.2 | -4.67 | 358.54 | 1,614.63 | 6,538.16 | 23,253.96 | 0.14 | 1,214.09 | 5,353.14 | -2.41 | 13.24 | 87.37 | 5,852.1 | 5,260.89 | 25,456.31 | 3.08 | 1,487.84 | 693.26 | 14,734.11 | -19.86 | 3,396.61 | 179.96 | 38,728.48 | 15.89 | 27,346.44 | 2,340.6 | 3,739.56 | — | |
| PPOFrames=40M, Trials=32018.07 | 1,115.94 | 413.46 | 2,168.93 | 2,102.1 | 1,470.46 | 596,807.27 | 643.94 | 9,387.77 | 1,460.59 | 39.41 | 78.61 | 124.98 | 3,344.63 | 3,106.14 | 90,169.6 | 7,180.43 | — | 321.2 | -27.51 | 15.87 | 267.73 | 1,196.2 | 509.81 | -5.5 | 394.45 | 2,199.74 | 7,195.24 | 23,283.31 | 0.74 | 1,482.77 | 5,199.37 | -5.81 | 12.83 | 52.76 | 6,744.13 | 5,487.17 | 24,688.07 | 8.65 | 1,120.15 | 632.17 | 13,643.8 | -21.8 | 4,410.87 | 152.72 | 43,208.99 | 14.66 | 27,549.55 | 2,743.4 | 1,813.9 | — | |
| BASELINEFrames=40M, Trials=32018.07 | 851.13 | 295.91 | 1,159.5 | 1,884.68 | 1,477.71 | 192,798.74 | 859.25 | 11,674.3 | 3,321.25 | 33.9 | 27.92 | 29.99 | 3,042.48 | 1,780.38 | 69,258.31 | 9,814.42 | — | 92.59 | -81.9 | 15.93 | 270.42 | 900.74 | 342.74 | -4.61 | 380.91 | 937.98 | 4,760.66 | 19,637.58 | 0.22 | 860.63 | 4,562.32 | -31.27 | 7.2 | 56.7 | 1,760.92 | 5,220.64 | 20,385.91 | 13.89 | 1,112.51 | 552.5 | 16,288.35 | -21.84 | 4,718.46 | 103.95 | 5,430.22 | 0.57 | 23,998.09 | 2,409.94 | 256.78 | — | |
| LSTMEnvironmental steps=50 M, Number of parameters=1.6 M2021.06 | 846 | 233 | 7,551 | — | — | — | 327 | 6,327 | — | — | — | — | — | — | — | 31,447 | — | — | — | — | — | 13,765 | — | — | 3,020 | — | — | 15,216 | — | 1,676 | — | — | 21 | — | 3,905 | 13,638 | — | 7.6 | 708 | 12,461 | 8,337 | — | — | — | 155,847 | — | — | — | 2,737 | 474 | |
| TRPOFrames=40M, Trials=32018.07 | 841.08 | 169.12 | 971.78 | 1,342.83 | 1,760.73 | 174,394.94 | 831.95 | 12,918.39 | 2,431.63 | 30.99 | 23.07 | 26.56 | 3,142.22 | 1,595.82 | 63,189.78 | 6,204.68 | — | 140.67 | -81.97 | 17.33 | 270.57 | 875.93 | 317.86 | -5.21 | 519.01 | 566.85 | 3,861.87 | 18,293.12 | 0.12 | 864.84 | 4,504.67 | -33.93 | -2.91 | 98.79 | 1,679.03 | 4,549.22 | 16,269.4 | 14.57 | 848.47 | 483.48 | 13,341.23 | -21.04 | 4,544.68 | 109.18 | 7,085.02 | 0.03 | 23,705.39 | 2,045.17 | 1,521.28 | — |