Predicting positive transfer for improved low-resource speech recognition using acoustic pseudo-tokens

Nay San, Georgios Paraskevopoulos, Aryaman Arora, Xiluo He, Prabhjot Kaur, Oliver Adams, Dan Jurafsky
Electrical Engineering and Systems Science, Audio and Speech Processing, Audio and Speech Processing (eess.AS), Computation and Language (cs.CL)
2024-02-03 00:00:00
While massively multilingual speech models like wav2vec 2.0 XLSR-128 can be directly fine-tuned for automatic speech recognition (ASR), downstream performance can still be relatively poor on languages that are under-represented in the pre-training data. Continued pre-training on 70-200 hours of untranscribed speech in these languages can help -- but what about languages without that much recorded data? For such cases, we show that supplementing the target language with data from a similar, higher-resource 'donor' language can help. For example, continued pre-training on only 10 hours of low-resource Punjabi supplemented with 60 hours of donor Hindi is almost as good as continued pretraining on 70 hours of Punjabi. By contrast, sourcing data from less similar donors like Bengali does not improve ASR performance. To inform donor language selection, we propose a novel similarity metric based on the sequence distribution of induced acoustic units: the Acoustic Token Distribution Similarity (ATDS). Across a set of typologically different target languages (Punjabi, Galician, Iban, Setswana), we show that the ATDS between the target language and its candidate donors precisely predicts target language ASR performance.
PDF: Predicting positive transfer for improved low-resource speech recognition using acoustic pseudo-tokens.pdf
Empowered by ChatGPT