Automatic Speech Recognition on CHiME-4 (dev-simu)
5.6WER (%)W2T-E
Evaluation Results
| Method | Links | |
|---|---|---|
| W2T-ESystem=+ W2T-E, Backend Architecture=Token ASR (AED)2026.02 | 5.6 | |
| Whisper (real label)Adaptation strategy=Finetuning, Supervision type=Supervised2024.05 | 5.9 | |
| W2W-E (TF-GridNet)System=+ W2W-E (TF-GridNet), Backend Architecture=Continuous ASR (AED)2026.02 | 5.9 | |
| STARAdaptation strategy=STAR, Supervision type=Unsupervised2024.05 | 6.4 | |
| UTTfilterAdaptation strategy=Self-training, Filtering=Utterance-level, Supervision type=Unsupervised2024.05 | 6.6 | |
| TOKreweight (A_t)Adaptation strategy=Self-training, Token-level re-weighting=A_t, Supervision type=Unsupervised2024.05 | 6.6 | |
| W2T-ESystem=+ W2T-E, Backend Architecture=Token ASR (CTC-only)2026.02 | 6.6 | |
| TOKreweight (G_t)Adaptation strategy=Self-training, Token-level re-weighting=G_t, Supervision type=Unsupervised2024.05 | 6.7 | |
| IRTFpost-filter=none2019.05 | 6.98 | |
| IRTFpost-filter=Wiener2019.05 | 6.98 | |
| Whisper (frozen)Adaptation strategy=None, Supervision type=Zero-shot2024.05 | 7 | |
| Whisper (self-train.)Adaptation strategy=Self-training, Supervision type=Unsupervised2024.05 | 7 | |
| MVDRpost-filter=Wiener2019.05 | 7.5 | |
| GEVpost-filter=BAN2019.05 | 7.78 | |
| MVDRpost-filter=none2019.05 | 7.95 | |
| WavLM (Weighted-sum)System=WavLM (Weighted-sum), Backend Architecture=Continuous ASR (AED)2026.02 | 8.1 | |
| Unprocessedpost-filter=none2019.05 | 8.86 | |
| W2W-E (Conv-TasNet)System=+ W2W-E (Conv-TasNet), Backend Architecture=Continuous ASR (AED)2026.02 | 9.1 | |
| W2W-E (TF-GridNet)System=+ W2W-E (TF-GridNet), Backend Architecture=Token ASR (AED)2026.02 | 9.2 | |
| GEVpost-filter=none2019.05 | 9.49 | |
| V2T-E (E-Branchformer)System=+ V2T-E (E-Branchformer), Backend Architecture=Token ASR (AED)2026.02 | 9.8 | |
| V2T-E (TCN)System=+ V2T-E (TCN), Backend Architecture=Token ASR (AED)2026.02 | 10.4 | |
| V2T-E (MLP)System=+ V2T-E (MLP), Backend Architecture=Token ASR (AED)2026.02 | 11.5 | |
| W2W-E (Conv-TasNet)System=+ W2W-E (Conv-TasNet), Backend Architecture=Token ASR (AED)2026.02 | 12.9 | |
| T2T-ESystem=+ T2T-E, Backend Architecture=Token ASR (AED)2026.02 | 17 | |
| WavLMSystem=WavLM, Backend Architecture=Token ASR (AED)2026.02 | 17.3 | |
| FBANKSystem=FBANK, Backend Architecture=Continuous ASR (AED)2026.02 | 18 | |
| WavLMSystem=WavLM, Backend Architecture=Token ASR (CTC-only)2026.02 | 21.9 |