Domain Adversarial Neural Networks for Dysarthric Speech Recognition
Preprint
Abstract
Speech recognition systems have improved dramatically over the last few years, however, their performance is significantly degraded for the cases of accented or impaired speech. This work explores domain adversarial neural networks (DANN) for speaker-independent speech recognition on the UAS dataset of dysarthric speech. The classification task on 10 spoken digits is performed using an end-to-end CNN taking raw audio as input. The results are compared to a speaker-adaptive (SA) model as well as speaker-dependent (SD) and multi-task learning models (MTL). The experiments conducted in this paper show that DANN achieves an absolute recognition rate of 74.91% and outperforms the baseline by 12.18%. Additionally, the DANN model achieves comparable results to the SA model's recognition rate of 77.65%. We also observe that when labelled dysarthric speech data is available DANN and MTL perform similarly, but when they are not DANN performs better than MTL.Keywords
All Related Versions
This publication has 8 references indexed in Scilit:
- Deep Learning Approaches for Understanding Simple Speech Commands2020
- Personalizing ASR for Dysarthric and Accented Speech with Limited Data2019
- Improved Accented Speech Recognition Using Accent Embeddings and Multi-task Learning2018
- Domain Adaptation Using Factorized Hidden Layer for Robust Automatic Speech Recognition2018
- Data Augmentation Using Healthy Speech for Dysarthric Speech Recognition2018
- A Multi-Views Multi-Learners Approach Towards Dysarthric Speech Recognition Using Multi-Nets Artificial Neural NetworksIEEE Transactions on Neural Systems and Rehabilitation Engineering, 2014
- Dysarthric speech recognition using dysarthria-severity-dependent and speaker-adaptive models2013
- The TORGO database of acoustic and articulatory speech from speakers with dysarthriaLanguage Resources and Evaluation, 2011