free-0c31fbbc
done · peak unseen_mismatch 92.5
Referenced by
α schedule
α 5.0 ×10 α 4.7 ×10 α 4.9 ×10 α 4.6 ×10 α 4.8 ×10 α 4.4 ×10 α 4.9 ×10 α 4.8 ×10 α 4.5 ×10 α 4.6 ×10 α 4.7 ×20 α 4.3 ×10 α 4.9 ×10 α 5.0 ×10 α 4.9 ×10 α 0.8 ×10 α 0.9 ×10 α 0.7 ×10 α 0.8 ×10 α 0.7 ×10 α 1.1 ×10 α 0.8 ×10 α 1.5 ×10 α 0.9 ×10 α 0.3 ×10 α 0.5 ×10 α 1.1 ×10 α 0.6 ×10 α 1.0 ×110
Figures
Results
accuracy by training iteration, mean ± sd over seeds · schedule ends at α = 1.0
seen_* probed at reference α = 1.4 · unseen_* = held-out pairings, uniform (α-independent) · 1000 pairs per condition
| iteration | seen_match | seen_mismatch | unseen_match | unseen_mismatch |
|---|---|---|---|---|
| 100 | 61.0 ± 7.7 | 57.4 ± 6.4 | 54.4 ± 5.8 | 51.8 ± 3.7 |
| 110 | 65.6 ± 9.2 | 61.9 ± 8.2 | 56.5 ± 7.7 | 53.8 ± 5.4 |
| 120 | 70.5 ± 9.5 | 66.9 ± 8.5 | 59.6 ± 8.4 | 56.3 ± 6.2 |
| 130 | 74.9 ± 8.7 | 71.5 ± 7.8 | 62.1 ± 8.9 | 58.7 ± 6.8 |
| 140 | 77.5 ± 7.3 | 74.7 ± 6.6 | 63.4 ± 8.4 | 60.2 ± 6.5 |
| 150 | 78.7 ± 6.1 | 76.1 ± 5.4 | 63.6 ± 7.5 | 60.7 ± 5.8 |
| 160 | 79.7 ± 5.1 | 77.3 ± 4.5 | 63.8 ± 6.8 | 60.9 ± 5.6 |
| 170 | 83.2 ± 6.5 | 80.8 ± 6.2 | 69.8 ± 9.1 | 66.5 ± 8.1 |
| 180 | 90.7 ± 7.0 | 88.0 ± 7.5 | 83.0 ± 10.7 | 79.1 ± 10.8 |
| 190 | 95.4 ± 6.3 | 93.0 ± 7.1 | 91.9 ± 8.8 | 88.3 ± 9.3 |
| 200 | 97.3 ± 5.1 | 94.6 ± 6.4 | 95.1 ± 6.9 | 90.7 ± 8.0 |
| 210 | 98.0 ± 4.1 | 90.8 ± 10.6 | 96.7 ± 5.4 | 88.1 ± 10.6 |
| 220 | 99.1 ± 1.2 | 87.3 ± 11.5 | 98.3 ± 2.0 | 84.6 ± 12.3 |
| 230 | 100.0 ± 0.0 | 82.8 ± 15.3 | 99.9 ± 0.2 | 80.4 ± 14.8 |
| 240 | 100.0 ± 0.0 | 81.3 ± 13.4 | 100.0 ± 0.1 | 80.0 ± 13.4 |
| 250 | 100.0 ± 0.0 | 90.3 ± 11.5 | 100.0 ± 0.0 | 88.8 ± 12.2 |
| 260 | 100.0 ± 0.0 | 72.0 ± 4.4 | 100.0 ± 0.0 | 70.5 ± 1.2 |
| 270 | 100.0 ± 0.0 | 63.9 ± 4.3 | 100.0 ± 0.0 | 63.8 ± 1.6 |
| 280 | 100.0 ± 0.0 | 59.0 ± 4.0 | 100.0 ± 0.0 | 59.9 ± 2.0 |
| 290 | 100.0 ± 0.0 | 71.7 ± 4.0 | 100.0 ± 0.0 | 73.0 ± 2.6 |
| 300 | 100.0 ± 0.0 | 81.2 ± 13.3 | 100.0 ± 0.0 | 82.3 ± 12.8 |
| 310 | 100.0 ± 0.0 | 73.0 ± 0.0 | 100.0 ± 0.0 | 72.4 ± 0.0 |
| 320 | 100.0 ± 0.0 | 93.0 ± 0.0 | 100.0 ± 0.0 | 92.5 ± 0.0 |