free-9a4f8259
done · peak unseen_mismatch 95.5
Referenced by
α schedule
α 5.0 ×10 α 4.8 ×10 α 4.9 ×10 α 5.0 ×10 α 4.1 ×20 α 4.2 ×10 α 4.4 ×10 α 4.6 ×10 α 4.8 ×10 α 4.4 ×10 α 4.7 ×10 α 4.9 ×10 α 4.1 ×10 α 4.8 ×10 α 5.0 ×20 α 2.4 ×10 α 0.0 ×20 α 0.4 ×10 α 0.0 ×10 α 3.0 ×10 α 4.8 ×10 α 4.6 ×10 α 3.5 ×10 α 2.3 ×10 α 1.0 ×10 α 0.4 ×10 α 0.5 ×110
Figures
Results
accuracy by training iteration, mean ± sd over seeds · schedule ends at α = 0.5
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 | 61.4 ± 7.9 | 57.8 ± 6.7 | 54.1 ± 5.9 | 51.8 ± 3.2 |
| 110 | 66.6 ± 9.2 | 62.9 ± 8.1 | 56.7 ± 7.5 | 54.1 ± 5.2 |
| 120 | 71.4 ± 9.2 | 67.8 ± 8.6 | 59.6 ± 8.4 | 56.6 ± 6.3 |
| 130 | 75.0 ± 7.8 | 71.7 ± 7.6 | 61.4 ± 8.1 | 58.4 ± 6.5 |
| 140 | 77.7 ± 6.5 | 74.8 ± 6.0 | 63.1 ± 7.6 | 60.1 ± 6.0 |
| 150 | 79.3 ± 5.0 | 76.8 ± 4.6 | 64.0 ± 7.0 | 61.1 ± 5.9 |
| 160 | 79.9 ± 4.2 | 77.5 ± 3.7 | 63.4 ± 6.5 | 60.9 ± 5.7 |
| 170 | 80.3 ± 3.7 | 78.1 ± 3.2 | 63.5 ± 6.0 | 61.2 ± 5.1 |
| 180 | 82.2 ± 3.5 | 80.2 ± 3.2 | 65.2 ± 6.2 | 62.9 ± 5.4 |
| 190 | 86.2 ± 5.5 | 84.1 ± 5.2 | 73.6 ± 9.2 | 70.6 ± 8.5 |
| 200 | 92.8 ± 6.7 | 90.7 ± 6.9 | 87.7 ± 9.6 | 84.4 ± 9.8 |
| 210 | 97.6 ± 4.0 | 95.6 ± 5.6 | 95.9 ± 6.0 | 92.9 ± 7.7 |
| 220 | 98.9 ± 1.5 | 93.8 ± 9.6 | 98.2 ± 2.5 | 91.4 ± 10.8 |
| 230 | 99.4 ± 0.8 | 90.0 ± 14.5 | 98.6 ± 1.7 | 87.2 ± 15.2 |
| 240 | 98.8 ± 0.5 | 95.2 ± 1.9 | 96.8 ± 2.0 | 89.7 ± 5.0 |
| 250 | 91.2 ± 2.5 | 88.3 ± 1.5 | 78.2 ± 7.7 | 74.5 ± 5.5 |
| 260 | 88.2 ± 0.7 | 86.6 ± 0.9 | 75.7 ± 3.2 | 73.5 ± 3.0 |
| 270 | 94.2 ± 0.7 | 92.1 ± 0.4 | 86.2 ± 0.3 | 82.1 ± 0.8 |
| 280 | 99.4 ± 0.3 | 98.2 ± 0.9 | 98.8 ± 1.1 | 95.5 ± 2.6 |