Chi tiết| Tên | Multilingual Shifting Deep Bottleneck Features for low-resource ASR |
| Lĩnh vực | Tin học |
| Tác giả | Quoc Bao Nguyen, Jonas Gehring, Markus Muller ,Sebastian Stuker, Alexander Waibel |
| Nhà xuất bản / Tạp chí | |
| Số hiệu ISSN/ISBN | |
| Tóm tắt nội dung | |
In this work, we propose a deep bottleneck feature architecture that is able to leverage data from multiple languages. Wealso show that tonal features are helpful for non-tonal languages. Evaluations are performed on a low-resource conversational telephone speech transcription task in Bengali, whileadditional data for DBNF training is provided in Assamese,Pashto, Tagalog, Turkish, and Vietnamese. We obtain relativereductions of up to 17.3% and 9.4% WER over mono-lingualGMMs and DBNFs, respectively.
|
|
