Adversarial Multilingual Training for Low-Resource Speech Recognition

Jiangyan Yi; Jianhua Tao; Zhengqi Wen; Ye Bai

ICASSP 2018accepted0 citations

Adversarial Multilingual Training for Low-Resource Speech Recognition

Jiangyan Yi, Jianhua Tao, Zhengqi Wen, Ye Bai

Abstract

This paper proposes an adversarial multilingual training to train bottleneck (BN) networks for the target language. A parallel shared-exclusive model is also proposed to train the BN network. Adversarial training is used to ensure that the shared layers can learn language-invariant features. Experiments are conducted on IARPA Babel datasets. The results show that the proposed adversarial multilingual BN model outperforms the baseline BN model by up to 8.9% relative word error rate (WER) reduction. The results also show that the proposed parallel shared-exclusive model achieves up to 1.7% relative WER reduction when compared with the stacked share-exclusive model.

BibTeX

@inproceedings{icassp2018_adversarialmulti,
  title = {Adversarial Multilingual Training for Low-Resource Speech Recognition},
  author = {Jiangyan Yi and Jianhua Tao and Zhengqi Wen and Ye Bai},
  booktitle = {ICASSP 2018},
  year = {2018}
}