This is the latest version of this eprint.
Farooq, M.U. and Hain, T. orcid.org/0000-0003-0939-3464 (2023) Learning cross-lingual mappings for data augmentation to improve low-resource speech recognition. In: Interspeech 2023 Proceedings. Interspeech 2023, 20-24 Aug 2023, Dublin, Ireland. International Speech Communication Association , pp. 5072-5076.
Abstract
Exploiting cross-lingual resources is an effective way to compensate for data scarcity of low resource languages. Recently, a novel multilingual model fusion technique has been proposed where a model is trained to learn cross-lingual acoustic-phonetic similarities as a mapping function. However, handcrafted lexicons have been used to train hybrid DNN-HMM ASR systems. To remove this dependency, we extend the concept of learnable cross-lingual mappings for end-to-end speech recognition. Furthermore, mapping models are employed to transliterate the source languages to the target language without using parallel data. Finally, the source audio and its transliteration is used for data augmentation to retrain the target language ASR. The results show that any source language ASR model can be used for a low-resource target language recognition followed by proposed mapping model. Furthermore, data augmentation results in a relative gain up to 5% over baseline monolingual model.
Metadata
Item Type: | Proceedings Paper |
---|---|
Authors/Creators: |
|
Copyright, Publisher and Additional Information: | © 2023 ISCA. Reproduced in accordance with the publisher's self-archiving policy. |
Keywords: | automatic speech recognition; low-resource; cross-lingual; multilingual; data augmentation |
Dates: |
|
Institution: | The University of Sheffield |
Academic Units: | The University of Sheffield > Faculty of Engineering (Sheffield) > Department of Computer Science (Sheffield) |
Depositing User: | Symplectic Sheffield |
Date Deposited: | 02 Aug 2023 12:09 |
Last Modified: | 15 Sep 2023 14:12 |
Status: | Published |
Publisher: | International Speech Communication Association |
Refereed: | Yes |
Identification Number: | 10.21437/Interspeech.2023-1613 |
Related URLs: | |
Open Archives Initiative ID (OAI ID): | oai:eprints.whiterose.ac.uk:202071 |
Available Versions of this Item
-
Learning cross-lingual mappings for data augmentation to improve low-resource speech recognition. (deposited 02 Aug 2023 10:35)
- Learning cross-lingual mappings for data augmentation to improve low-resource speech recognition. (deposited 02 Aug 2023 12:09) [Currently Displayed]