Classification on Warehouse Domain (test)
87.48Macro F1Active Learning + Explanation
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Active Learning + ExplanationAL Rounds=3, Num. Manual Labels=91, Accum. Manual Labels=1246, No. Pseudo Labels=2754, Training Samples=4000, Annotation Saved (%)=68.852026.05 | 87.48 | 89.6 | |
| Active Learning + ExplanationAL Rounds=2, Num. Manual Labels=55, Accum. Manual Labels=1155, No. Pseudo Labels=1845, Training Samples=3000, Annotation Saved (%)=61.502026.05 | 86.67 | 88.8 | |
| Active Learning + ExplanationAL Rounds=0, Num. Manual Labels=961, Accum. Manual Labels=961, No. Pseudo Labels=539, Training Samples=1500, Annotation Saved (%)=35.932026.05 | 84.02 | 87 | |
| Active Learning + ExplanationAL Rounds=1, Num. Manual Labels=139, Accum. Manual Labels=1100, No. Pseudo Labels=900, Training Samples=2000, Annotation Saved (%)=45.002026.05 | 82.91 | 85.2 | |
| Active LearningAL Rounds=0, Num. Manual Labels=961, Accum. Manual Labels=961, No. Pseudo Labels=539, Training Samples=1500, Annotation Saved (%)=35.932026.05 | 77.34 | 81.8 | |
| Fine-tunedAL Rounds=-, Num. Manual Labels=4000, Accum. Manual Labels=4000, No. Pseudo Labels=0, Training Samples=4000, Annotation Saved (%)=02026.05 | 75.03 | 83.2 | |
| Zero-shot + ExplanationAL Rounds=-, Num. Manual Labels=0, Accum. Manual Labels=0, No. Pseudo Labels=0, Training Samples=0, Annotation Saved (%)=02026.05 | 44.11 | 50.6 | |
| Zero-shotAL Rounds=-, Num. Manual Labels=0, Accum. Manual Labels=0, No. Pseudo Labels=0, Training Samples=0, Annotation Saved (%)=02026.05 | 34.21 | 39 |