Active Learning on Synthetic phase-competition system_seed = 12 (test)
1.945NLL (Mean)MI-LB (MaxDist)
Evaluation results
| Method | Links | ||
|---|---|---|---|
| MI-LB (MaxDist)Training-set size (n)=550, Number of seeds=5, system_seed=12, Selection Strategy=MaxDist, MaxDist score weight (w)=12026.05 | 1.945 | 1.87 | |
| MI-LB (top-k)Training-set size (n)=550, Number of seeds=5, system_seed=12, Selection Strategy=top-k2026.05 | 1.985 | 1.911 | |
| Core-SetTraining-set size (n)=550, Number of seeds=5, system_seed=122026.05 | 1.989 | 1.951 | |
| Epistemic Variance (SBAL)Training-set size (n)=550, Number of seeds=5, system_seed=12, Selection Strategy=SBAL, SBAL temperature (T)=1.02026.05 | 2.029 | 1.977 | |
| Epistemic Variance (top-k)Training-set size (n)=550, Number of seeds=5, system_seed=12, Selection Strategy=top-k2026.05 | 2.062 | 1.918 | |
| MI-LB (SBAL)Training-set size (n)=550, Number of seeds=5, system_seed=12, Selection Strategy=SBAL, SBAL temperature (T)=1.02026.05 | 2.109 | 2.057 | |
| BAITTraining-set size (n)=550, Number of seeds=5, system_seed=122026.05 | 2.19 | 2.087 | |
| RandomTraining-set size (n)=550, Number of seeds=5, system_seed=122026.05 | 2.209 | 2.136 |