free-5411cbfe
done · peak unseen_mismatch 93.5
Referenced by
α schedule
α 4.761 ×160 α 0.8109 ×140 α 0.8109 ×100
Figures
Results
accuracy by training iteration, mean ± sd over seeds · schedule ends at α = 0.8109
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 | 61.9 ± 7.4 | 58.2 ± 6.2 | 54.1 ± 5.8 | 51.7 ± 3.4 |
| 110 | 67.3 ± 8.6 | 63.4 ± 8.1 | 56.5 ± 7.2 | 53.6 ± 5.2 |
| 120 | 72.1 ± 8.6 | 68.5 ± 8.2 | 59.1 ± 8.1 | 56.2 ± 6.4 |
| 130 | 75.6 ± 7.5 | 72.6 ± 7.3 | 61.1 ± 8.1 | 58.2 ± 6.6 |
| 140 | 77.8 ± 6.2 | 75.1 ± 5.6 | 62.1 ± 7.3 | 59.4 ± 6.2 |
| 150 | 78.9 ± 5.1 | 76.5 ± 4.5 | 62.4 ± 6.9 | 59.9 ± 5.8 |
| 160 | 79.6 ± 4.3 | 77.4 ± 3.8 | 62.4 ± 6.2 | 60.2 ± 5.5 |
| 170 | 82.9 ± 5.6 | 80.4 ± 5.3 | 68.3 ± 8.4 | 65.4 ± 7.4 |
| 180 | 90.3 ± 6.1 | 87.8 ± 6.4 | 81.4 ± 9.9 | 77.8 ± 9.7 |
| 190 | 96.3 ± 4.3 | 94.3 ± 5.1 | 92.7 ± 6.7 | 89.4 ± 7.6 |
| 200 | 98.3 ± 3.0 | 96.4 ± 4.7 | 96.6 ± 4.2 | 93.5 ± 5.9 |
| 210 | 98.6 ± 2.1 | 93.3 ± 7.8 | 97.9 ± 2.6 | 90.7 ± 8.3 |
| 220 | 99.9 ± 0.2 | 88.7 ± 9.0 | 99.8 ± 0.3 | 86.8 ± 9.9 |
| 230 | 100.0 ± 0.0 | 88.9 ± 8.3 | 100.0 ± 0.0 | 89.0 ± 8.3 |
| 240 | 100.0 ± 0.0 | 91.7 ± 8.2 | 100.0 ± 0.0 | 91.9 ± 8.0 |
| 250 | 100.0 ± 0.0 | 70.7 ± 0.0 | 100.0 ± 0.0 | 69.8 ± 0.0 |
| 260 | 100.0 ± 0.0 | 75.0 ± 0.0 | 100.0 ± 0.0 | 73.5 ± 0.0 |
| 270 | 100.0 ± 0.0 | 60.8 ± 0.0 | 100.0 ± 0.0 | 61.9 ± 0.0 |
| 280 | 100.0 ± 0.0 | 55.6 ± 0.0 | 100.0 ± 0.0 | 57.4 ± 0.0 |
| 290 | 100.0 ± 0.0 | 49.9 ± 0.0 | 100.0 ± 0.0 | 49.9 ± 0.0 |
| 300 | 100.0 ± 0.0 | 46.8 ± 0.0 | 100.0 ± 0.0 | 47.0 ± 0.0 |
| 310 | 100.0 ± 0.0 | 47.7 ± 0.0 | 100.0 ± 0.0 | 47.8 ± 0.0 |
| 320 | 100.0 ± 0.0 | 50.0 ± 0.0 | 100.0 ± 0.0 | 49.9 ± 0.0 |
| 330 | 100.0 ± 0.0 | 53.6 ± 0.0 | 100.0 ± 0.0 | 53.1 ± 0.0 |
| 340 | 100.0 ± 0.0 | 49.9 ± 0.0 | 100.0 ± 0.0 | 49.9 ± 0.0 |
| 350 | 100.0 ± 0.0 | 49.7 ± 0.0 | 100.0 ± 0.0 | 49.4 ± 0.0 |
| 360 | 100.0 ± 0.0 | 47.3 ± 0.0 | 100.0 ± 0.0 | 45.0 ± 0.0 |
| 370 | 100.0 ± 0.0 | 52.5 ± 0.0 | 100.0 ± 0.0 | 54.7 ± 0.0 |
| 380 | 100.0 ± 0.0 | 52.1 ± 0.0 | 100.0 ± 0.0 | 53.2 ± 0.0 |
| 390 | 100.0 ± 0.0 | 52.9 ± 0.0 | 100.0 ± 0.0 | 54.0 ± 0.0 |
| 400 | 100.0 ± 0.0 | 54.6 ± 0.0 | 100.0 ± 0.0 | 55.3 ± 0.0 |