Acoustic Discriminability (ABX) on Zero Resource Speech Challenge (en, fr, zh, de, wo) 2017
3.81ABX Triphone 1s (WS)MAUBERT FEAT + MPR
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| MAUBERT FEAT + MPRLayer=11, # units=768, Evaluation Protocol=supervised FT (10 h)2025.12 | 3.81 | 4.26 | 2.86 | 3.25 | 3.28 | 3.71 | 3.53 | |
| MAUBERT PHONE + MPRLayer=12, # units=768, Evaluation Protocol=supervised FT (10 h)2025.12 | 3.92 | 4.61 | 2.57 | 3.08 | 2.86 | 3.32 | 3.39 | |
| HuBERT-base + MPRLayer=11, # units=768, Evaluation Protocol=supervised FT (10 h)2025.12 | 4.26 | 4.84 | 3.25 | 3.73 | 3.89 | 4.36 | 4.05 | |
| XEUSLayer=18, # units=1024, Evaluation Protocol=Zero-shot2025.12 | 4.67 | 5.68 | 4.19 | 4.91 | 4.29 | 4.99 | 4.79 | |
| MAUBERT-PHONE + K-means (proj)Layer=10, # units=768, Evaluation Protocol=self-supervised FT (10 h)2025.12 | 4.84 | 5.62 | 4.17 | 4.81 | 4.38 | 5.15 | 4.83 | |
| MAUBERT-FEAT + feat. freq.Layer=9, # units=768, Evaluation Protocol=self-supervised FT (10 h)2025.12 | 4.99 | 5.8 | 4.19 | 4.86 | 4.39 | 5.09 | 4.89 | |
| MAUBERT-FEAT + K-means (L9)Layer=10, # units=768, Evaluation Protocol=self-supervised FT (10 h)2025.12 | 5.01 | 5.56 | 4.19 | 4.71 | 4.38 | 5 | 4.81 | |
| MAUBERT-FEAT + phone freq.Layer=9, # units=768, Evaluation Protocol=self-supervised FT (10 h)2025.12 | 5.09 | 5.87 | 4.31 | 5.01 | 4.53 | 5.24 | 5.01 | |
| MAUBERT-PHONE + phone freq.Layer=10, # units=768, Evaluation Protocol=self-supervised FT (10 h)2025.12 | 5.11 | 5.79 | 4.09 | 4.72 | 4.24 | 4.86 | 4.8 | |
| MAUBERT-PHONE + all phonesLayer=10, # units=768, Evaluation Protocol=self-supervised FT (10 h)2025.12 | 5.15 | 5.89 | 4.05 | 4.7 | 4.2 | 4.83 | 4.8 | |
| MAUBERT-FEAT + K-means (feat)Layer=9, # units=768, Evaluation Protocol=self-supervised FT (10 h)2025.12 | 5.16 | 5.92 | 4.3 | 4.98 | 4.51 | 5.2 | 5.01 | |
| MAUBERT PHONELayer=proj, # units=1024, Evaluation Protocol=Zero-shot2025.12 | 5.36 | 6.44 | 4.68 | 5.58 | 4.68 | 5.6 | 5.39 | |
| HuBERT-base + PRLayer=ws, # units=768, Evaluation Protocol=supervised FT (10 h)2025.12 | 5.52 | 6.67 | 4.1 | 4.99 | 4.51 | 5.49 | 5.21 | |
| MAUBERT-PHONE + K-means (phone)Layer=10, # units=768, Evaluation Protocol=self-supervised FT (10 h)2025.12 | 5.52 | 6.16 | 4.14 | 4.76 | 4.28 | 4.86 | 4.95 | |
| HuBERT-base + K-means (L11)Layer=10, # units=768, Evaluation Protocol=self-supervised FT (10 h)2025.12 | 5.65 | 6.38 | 4.79 | 5.4 | 5.09 | 5.77 | 5.51 | |
| MAUBERT FEATLayer=9, # units=768, Evaluation Protocol=Zero-shot2025.12 | 5.86 | 6.84 | 4.78 | 5.57 | 4.86 | 5.68 | 5.6 | |
| WavLM-largeLayer=24, # units=1024, Evaluation Protocol=Zero-shot2025.12 | 5.87 | 6.82 | 5.26 | 5.93 | 5.15 | 5.86 | 5.82 | |
| WavLM-base+Layer=7, # units=768, Evaluation Protocol=Zero-shot2025.12 | 6.13 | 6.97 | 5.07 | 5.83 | 5.16 | 6.04 | 5.87 | |
| HuBERT-baseLayer=11, # units=768, Evaluation Protocol=Zero-shot2025.12 | 6.21 | 7.42 | 5.31 | 6.21 | 5.62 | 6.62 | 6.23 | |
| mHuBERT-147Layer=7, # units=768, Evaluation Protocol=Zero-shot2025.12 | 6.93 | 8.13 | 5.75 | 6.49 | 6.67 | 7.78 | 6.96 | |
| MMS-1BLayer=34, # units=1280, Evaluation Protocol=Zero-shot2025.12 | 7.58 | 9.02 | 6.91 | 7.91 | 6.91 | 7.83 | 7.69 | |
| MFCCLayer=-, # units=39, Evaluation Protocol=Zero-shot2025.12 | 14.78 | 25.58 | 14.7 | 25.33 | 14.7 | 25.32 | 20.07 |