imagining syntax

← efficiency_v1

ann_burst_hold10

done · peak unseen_mismatch 96.8

Referenced by

α schedule

α 3.0 ×150 α 1.0 ×150 α 2.1429 ×150 α 1.0 ×150 α 1.2857 ×150 α 1.0 ×150 α 1.0 ×150 α 1.0 ×150 α 1.0 ×150 α 1.0 ×150 α 1.0 ×150 α 1.0 ×150 α 1.0 ×150 α 1.0 ×150 α 1.0 ×150 α 1.0 ×150

direction
mixed
α range
1.0–3.0
stages
16

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
150 70.1 ± 16.162.2 ± 12.367.1 ± 15.660.3 ± 10.7
300 100.0 ± 0.040.7 ± 27.0100.0 ± 0.040.8 ± 27.3
450 99.8 ± 0.393.8 ± 15.199.5 ± 0.993.0 ± 14.8
600 100.0 ± 0.097.7 ± 3.1100.0 ± 0.096.8 ± 3.2
750 100.0 ± 0.097.2 ± 4.199.7 ± 0.594.6 ± 5.9
900 100.0 ± 0.098.1 ± 2.599.7 ± 0.994.6 ± 5.6
1050 100.0 ± 0.096.4 ± 3.099.7 ± 0.490.5 ± 7.7
1200 100.0 ± 0.090.1 ± 9.299.9 ± 0.279.4 ± 14.8
1350 100.0 ± 0.090.9 ± 8.499.6 ± 0.578.8 ± 14.7
1500 100.0 ± 0.087.7 ± 8.799.8 ± 0.475.0 ± 16.3
1650 100.0 ± 0.087.0 ± 13.398.9 ± 0.970.3 ± 18.7
1800 100.0 ± 0.083.5 ± 14.899.4 ± 0.668.2 ± 17.6
1950 100.0 ± 0.085.2 ± 12.098.5 ± 1.569.2 ± 16.6
2100 100.0 ± 0.075.9 ± 19.598.4 ± 2.059.6 ± 21.0
2250 100.0 ± 0.079.7 ± 10.798.5 ± 1.562.6 ± 16.4
2400 100.0 ± 0.079.5 ± 9.697.8 ± 2.261.2 ± 14.2

Downloads