free-c667dd31
done · peak unseen_mismatch 96.3
Referenced by
α schedule
α 4.9 ×10 α 4.8 ×10 α 4.6 ×10 α 4.3 ×10 α 5.0 ×20 α 4.4 ×10 α 4.8 ×10 α 4.5 ×10 α 4.8 ×10 α 5.0 ×10 α 4.8 ×10 α 5.0 ×10 α 4.7 ×10 α 4.8 ×10 α 4.7 ×10 α 0.5 ×10 α 1.1 ×10 α 0.3 ×10 α 0.9 ×10 α 0.3 ×10 α 1.1 ×10 α 1.0 ×10 α 0.6 ×10 α 1.1 ×10 α 0.9 ×10 α 0.8 ×10 α 1.1 ×10 α 0.8 ×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 | 60.5 ± 7.5 | 57.0 ± 6.5 | 53.9 ± 5.5 | 51.8 ± 3.6 |
| 110 | 65.9 ± 8.8 | 62.1 ± 7.8 | 56.7 ± 6.8 | 53.8 ± 4.8 |
| 120 | 71.1 ± 9.0 | 67.3 ± 8.1 | 59.6 ± 8.1 | 56.4 ± 5.8 |
| 130 | 75.1 ± 8.1 | 71.8 ± 7.1 | 61.9 ± 8.4 | 58.5 ± 6.4 |
| 140 | 77.5 ± 6.7 | 74.7 ± 5.9 | 63.0 ± 7.5 | 60.2 ± 6.2 |
| 150 | 78.8 ± 5.1 | 76.2 ± 4.4 | 63.3 ± 7.0 | 60.6 ± 5.7 |
| 160 | 79.6 ± 4.6 | 77.3 ± 3.9 | 63.4 ± 6.7 | 60.9 ± 5.7 |
| 170 | 82.8 ± 6.2 | 80.2 ± 5.9 | 69.3 ± 9.1 | 66.1 ± 8.1 |
| 180 | 90.9 ± 6.5 | 88.4 ± 6.8 | 83.5 ± 10.3 | 79.6 ± 10.3 |
| 190 | 95.8 ± 5.9 | 93.3 ± 6.9 | 92.5 ± 8.1 | 88.8 ± 9.1 |
| 200 | 97.7 ± 5.2 | 93.8 ± 7.5 | 95.9 ± 6.8 | 90.3 ± 8.8 |
| 210 | 98.2 ± 6.5 | 88.6 ± 15.2 | 97.4 ± 8.0 | 85.6 ± 15.6 |
| 220 | 98.3 ± 5.3 | 88.6 ± 8.2 | 97.5 ± 7.2 | 85.4 ± 9.9 |
| 230 | 98.9 ± 2.7 | 88.5 ± 9.5 | 98.2 ± 4.3 | 87.0 ± 9.8 |
| 240 | 99.8 ± 0.2 | 88.8 ± 7.7 | 99.9 ± 0.2 | 87.5 ± 7.3 |
| 250 | 100.0 ± 0.0 | 76.3 ± 15.5 | 100.0 ± 0.0 | 76.8 ± 14.1 |
| 260 | 100.0 ± 0.0 | 85.3 ± 0.0 | 100.0 ± 0.0 | 84.2 ± 0.0 |
| 270 | 100.0 ± 0.0 | 86.2 ± 0.0 | 100.0 ± 0.0 | 86.2 ± 0.0 |
| 280 | 100.0 ± 0.0 | 76.3 ± 0.0 | 100.0 ± 0.0 | 77.4 ± 0.0 |
| 290 | 100.0 ± 0.0 | 87.1 ± 0.0 | 100.0 ± 0.0 | 85.2 ± 0.0 |
| 300 | 100.0 ± 0.0 | 95.6 ± 0.0 | 100.0 ± 0.0 | 96.3 ± 0.0 |