imagining syntax

← deform_v1

end_1.6

done · peak unseen_mismatch 97.0

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 1.0 ×19 α 1.0 ×19 α 1.0 ×19 α 1.0 ×19 α 1.0 ×19 α 1.0 ×19 α 1.0 ×19 α 1.0 ×19 α 2.3 ×19 α 2.3 ×19 α 2.3 ×19 α 2.3 ×19 α 2.3 ×19 α 2.3 ×19 α 2.3 ×19 α 2.3 ×19 α 1.0 ×19 α 1.0 ×19 α 1.0 ×19 α 1.0 ×19 α 1.0 ×19 α 1.0 ×19 α 1.0 ×19 α 1.0 ×19 α 1.6 ×19 α 1.6 ×19 α 1.6 ×19 α 1.6 ×19 α 1.6 ×18 α 1.6 ×18 α 1.6 ×18 α 1.6 ×18 α 1.0 ×18 α 1.0 ×18 α 1.0 ×18 α 1.0 ×18 α 1.0 ×18 α 1.0 ×18 α 1.0 ×18 α 1.0 ×18

direction
mixed
α range
1.0–3.0
stages
48

Figures

schedule.png
schedule.png
trajectory.png
trajectory.png

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

iterationseen_matchseen_mismatchunseen_matchunseen_mismatch
25 51.4 ± 1.549.5 ± 1.649.8 ± 1.949.7 ± 2.1
50 50.7 ± 1.949.6 ± 1.450.1 ± 1.550.8 ± 1.1
75 51.9 ± 1.850.1 ± 1.950.5 ± 1.750.0 ± 1.9
100 53.8 ± 2.952.1 ± 2.951.5 ± 1.850.0 ± 1.5
125 65.8 ± 10.560.5 ± 8.159.9 ± 9.254.9 ± 5.3
150 79.3 ± 12.473.6 ± 12.872.2 ± 14.166.3 ± 12.7
175 92.1 ± 12.983.0 ± 15.889.8 ± 14.679.1 ± 16.2
200 97.0 ± 11.280.7 ± 22.797.0 ± 10.678.2 ± 21.7
225 97.3 ± 11.485.9 ± 18.997.3 ± 11.084.2 ± 18.6
250 97.4 ± 11.186.7 ± 22.297.4 ± 10.485.2 ± 22.3
275 97.5 ± 10.688.5 ± 21.497.4 ± 10.886.7 ± 21.2
300 97.6 ± 10.690.1 ± 18.297.5 ± 10.788.3 ± 18.4
325 98.4 ± 2.890.7 ± 17.496.4 ± 3.786.1 ± 16.7
350 98.6 ± 1.195.0 ± 7.395.6 ± 3.789.7 ± 8.1
375 97.9 ± 1.495.8 ± 2.894.7 ± 4.089.4 ± 6.3
400 97.6 ± 1.695.6 ± 2.492.6 ± 4.687.8 ± 6.2
450 96.4 ± 1.694.6 ± 1.789.7 ± 4.884.5 ± 5.8
500 99.9 ± 0.199.3 ± 0.999.3 ± 1.097.0 ± 2.4
550 99.9 ± 0.199.4 ± 0.599.3 ± 0.896.1 ± 3.1
600 100.0 ± 0.199.3 ± 1.099.3 ± 0.894.9 ± 6.0
650 99.4 ± 0.497.8 ± 1.495.5 ± 2.688.7 ± 7.2
700 99.1 ± 0.597.9 ± 1.194.8 ± 2.487.8 ± 7.1
750 99.1 ± 0.597.8 ± 1.294.0 ± 2.387.6 ± 5.6
800 100.0 ± 0.199.5 ± 0.699.2 ± 0.795.3 ± 5.5
850 100.0 ± 0.199.4 ± 0.799.4 ± 0.794.2 ± 5.6
900 100.0 ± 0.199.3 ± 0.999.3 ± 0.692.6 ± 7.4

Downloads