https://scholars.lib.ntu.edu.tw/handle/123456789/640018
標題: | Prompting and Adapter Tuning For Self-Supervised Encoder-Decoder Speech Model | 作者: | Chang, Kai Wei Chen, Ming Hsin Lin, Yun Ping Hsu, Jing Neng Huang, Paul Kuo Ming Huang, Chien Yu Li, Shang Wen HUNG-YI LEE |
關鍵字: | adapter | automatic speech recognition | parameter-efficient tuning | Prompting | sequence generation | 公開日期: | 1-一月-2023 | 來源出版物: | 2023 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2023 | 摘要: | Prompting and adapter tuning have emerged as efficient alternatives to fine-tuning (FT) methods. However, existing studies on speech prompting focused on classification tasks and failed on more complex sequence generation tasks. Besides, adapter tuning is primarily applied with a focus on encoder-only self-supervised models. Our experiments show that prompting on Wav2Seq, a self-supervised encoder-decoder model, surpasses previous works in sequence generation tasks. It achieves a remarkable 53% relative improvement in word error rate for ASR and a 27% in F1 score for slot filling. Additionally, prompting competes with the FT method in the low-resource scenario. Moreover, we show the transferability of prompting and adapter tuning on Wav2Seq in cross-lingual ASR. When limited trainable parameters are involved, prompting and adapter tuning consistently outperform conventional FT across 7 languages. Notably, in the low-resource scenario, prompting consistently outperforms adapter tuning. |
URI: | https://scholars.lib.ntu.edu.tw/handle/123456789/640018 | ISBN: | 9798350306897 | DOI: | 10.1109/ASRU57964.2023.10389731 |
顯示於: | 電機工程學系 |
在 IR 系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。