v_0.7
done · peak unseen_mismatch 98.6
Referenced by
α schedule
α 3.0 ×19 α 3.0 ×19 α 3.0 ×19 α 3.0 ×19 α 3.0 ×19 α 3.0 ×19 α 3.0 ×19 α 3.0 ×19 α 0.7 ×19 α 0.7 ×19 α 0.7 ×19 α 0.7 ×19 α 0.7 ×19 α 0.7 ×19 α 0.7 ×19 α 0.7 ×19 α 2.1429 ×19 α 2.1429 ×19 α 2.1429 ×19 α 2.1429 ×19 α 2.1429 ×19 α 2.1429 ×19 α 2.1429 ×19 α 2.1429 ×19 α 0.7 ×19 α 0.7 ×19 α 0.7 ×19 α 0.7 ×19 α 0.7 ×19 α 0.7 ×19 α 0.7 ×19 α 0.7 ×19 α 1.2857 ×19 α 1.2857 ×19 α 1.2857 ×19 α 1.2857 ×19 α 1.2857 ×18 α 1.2857 ×18 α 1.2857 ×18 α 1.2857 ×18 α 0.7 ×18 α 0.7 ×18 α 0.7 ×18 α 0.7 ×18 α 0.7 ×18 α 0.7 ×18 α 0.7 ×18 α 0.7 ×18
Figures
Results
accuracy by training iteration, mean ± sd over seeds · schedule ends at α = 0.7
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 |
|---|---|---|---|---|
| 25 | 51.4 ± 1.5 | 49.5 ± 1.6 | 49.8 ± 1.9 | 49.7 ± 2.1 |
| 50 | 50.7 ± 1.9 | 49.6 ± 1.4 | 50.1 ± 1.5 | 50.8 ± 1.1 |
| 75 | 51.9 ± 1.8 | 50.1 ± 1.9 | 50.5 ± 1.7 | 50.0 ± 1.9 |
| 100 | 53.8 ± 2.9 | 52.1 ± 2.9 | 51.5 ± 1.8 | 50.0 ± 1.5 |
| 125 | 65.8 ± 10.5 | 60.5 ± 8.1 | 59.9 ± 9.2 | 54.9 ± 5.3 |
| 150 | 79.3 ± 12.4 | 73.6 ± 12.8 | 72.2 ± 14.1 | 66.3 ± 12.7 |
| 175 | 92.1 ± 12.8 | 79.9 ± 16.0 | 89.9 ± 14.1 | 76.8 ± 16.1 |
| 200 | 97.2 ± 11.1 | 74.1 ± 23.8 | 97.3 ± 10.5 | 72.4 ± 23.6 |
| 225 | 97.4 ± 11.4 | 74.0 ± 22.8 | 97.5 ± 11.0 | 72.3 ± 23.3 |
| 250 | 97.5 ± 11.0 | 73.9 ± 23.9 | 97.6 ± 10.4 | 72.8 ± 24.1 |
| 275 | 97.6 ± 10.5 | 66.6 ± 29.9 | 97.6 ± 10.6 | 66.2 ± 29.6 |
| 300 | 97.7 ± 10.1 | 72.9 ± 24.4 | 97.6 ± 10.3 | 71.5 ± 24.2 |
| 325 | 99.1 ± 3.7 | 82.0 ± 20.7 | 98.5 ± 4.0 | 79.8 ± 21.0 |
| 350 | 99.6 ± 0.5 | 90.3 ± 16.6 | 98.4 ± 2.3 | 86.6 ± 17.4 |
| 375 | 99.1 ± 0.9 | 96.3 ± 4.6 | 97.5 ± 2.4 | 92.7 ± 6.6 |
| 400 | 98.7 ± 1.3 | 96.2 ± 4.9 | 95.5 ± 4.6 | 90.7 ± 7.8 |
| 450 | 97.9 ± 1.5 | 96.5 ± 2.0 | 93.2 ± 4.7 | 88.9 ± 7.1 |
| 500 | 100.0 ± 0.0 | 99.3 ± 1.8 | 100.0 ± 0.1 | 98.6 ± 2.3 |
| 550 | 100.0 ± 0.0 | 97.7 ± 5.3 | 100.0 ± 0.1 | 95.7 ± 7.3 |
| 600 | 100.0 ± 0.0 | 94.8 ± 10.7 | 100.0 ± 0.1 | 92.0 ± 11.1 |
| 650 | 99.8 ± 0.2 | 96.8 ± 6.7 | 98.1 ± 1.5 | 91.0 ± 8.1 |
| 700 | 99.8 ± 0.2 | 98.2 ± 3.0 | 98.3 ± 1.1 | 92.5 ± 5.1 |
| 750 | 99.8 ± 0.2 | 98.9 ± 0.7 | 97.7 ± 1.5 | 92.2 ± 5.7 |
| 800 | 100.0 ± 0.0 | 99.2 ± 1.5 | 100.0 ± 0.0 | 97.1 ± 3.5 |
| 850 | 100.0 ± 0.0 | 98.5 ± 1.7 | 99.9 ± 0.3 | 93.4 ± 7.7 |
| 900 | 100.0 ± 0.0 | 95.3 ± 7.1 | 99.9 ± 0.4 | 87.5 ± 11.6 |