VPD screening on In-domain Clinical Recordings Standardized (held-out)
100AccuracyWhisper
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| WhisperClassifier=MLP2026.03 | 100 | 100 | 100 | 100 | |
| WhisperClassifier=XGBoost2026.03 | 100 | 100 | 100 | 100 | |
| WhisperClassifier=SVM2026.03 | 100 | 100 | 100 | 100 | |
| WhisperClassifier=Logistic Regression2026.03 | 100 | 100 | 100 | 100 | |
| HuBERTClassifier=MLP2026.03 | 100 | 100 | 100 | 100 | |
| HuBERTClassifier=XGBoost2026.03 | 100 | 100 | 100 | 100 | |
| SupCon NasalityEmbedding Dimension=256-d, Classifier=Logistic Regression, Chunk Size=0.5 s2026.03 | 100 | 100 | 100 | 100 | |
| SupCon NasalityEmbedding Dimension=256-d, Classifier=SVM, Chunk Size=0.5 s2026.03 | 100 | 100 | 100 | 100 | |
| SupCon NasalityEmbedding Dimension=256-d, Classifier=MLP, Chunk Size=0.5 s2026.03 | 100 | 100 | 100 | 100 | |
| SupCon NasalityEmbedding Dimension=256-d, Classifier=XGBoost, Chunk Size=0.5 s2026.03 | 100 | 100 | 100 | 100 | |
| HuBERTClassifier=SVM2026.03 | 99.2 | 92.9 | 99.6 | 96 | |
| Data2VecClassifier=XGBoost2026.03 | 99.2 | 92.9 | 99.6 | 96 | |
| MFCCClassifier=SVM2026.03 | 99.2 | 99.6 | 91.7 | 95.3 | |
| MFCCClassifier=MLP2026.03 | 99.2 | 99.6 | 91.7 | 95.3 | |
| MFCCClassifier=Logistic Regression2026.03 | 99.2 | 99.6 | 91.7 | 95.3 | |
| Data2VecClassifier=Logistic Regression2026.03 | 98.5 | 87.5 | 99.2 | 92.5 | |
| HuBERTClassifier=Logistic Regression2026.03 | 98.5 | 87.5 | 99.2 | 92.5 | |
| MFCCClassifier=XGBoost2026.03 | 98.5 | 91.3 | 91.3 | 91.3 | |
| Data2VecClassifier=SVM2026.03 | 96.9 | 80 | 98.4 | 86.7 | |
| Data2VecClassifier=MLP2026.03 | 95.4 | 75 | 97.6 | 82.1 | |
| Wav2Vec2Classifier=XGBoost2026.03 | 94.6 | 73.1 | 97.2 | 80.1 | |
| Wav2Vec2Classifier=SVM2026.03 | 94.6 | 73.1 | 97.2 | 80.1 | |
| Wav2Vec2Classifier=MLP2026.03 | 93 | 70 | 96.3 | 76.7 | |
| Wav2Vec2Classifier=Logistic Regression2026.03 | 90.7 | 66.7 | 95.1 | 72.4 |