Automated Speech Recognition on SlideSpeech Ag
10.8WERAdapted model (audio)
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Adapted model (audio)adaptation_type=audio2026.01 | 10.8 | 27.1 | |
| Adapted model (audio)Adaptation Modality=audio, Source Domain=DefinedAI, Tau (Specialization threshold)=0.642026.01 | 12.25 | 62.5 | |
| Text Denoising Adaptationadaptation_type=text2026.01 | 14.21 | 4.1 | |
| Ma et al.adaptation_type=text2026.01 | 14.23 | 4 | |
| Fang et al.adaptation_type=text2026.01 | 14.47 | 2.4 | |
| Base modeladaptation_type=none2026.01 | 14.82 | — | |
| Text Denoising AdaptationAdaptation Modality=text, Source Domain=DefinedAI, Tau (Specialization threshold)=0.642026.01 | 29.18 | 10.6 | |
| Ma et al.Adaptation Modality=text, Source Domain=DefinedAI, Tau (Specialization threshold)=0.642026.01 | 29.22 | 10.5 | |
| Fang et al.Adaptation Modality=text, Source Domain=DefinedAI, Tau (Specialization threshold)=0.642026.01 | 31.01 | 5 | |
| Base modelAdaptation Modality=none, Source Domain=DefinedAI, Tau (Specialization threshold)=0.642026.01 | 32.64 | — |