Publications

CONFERENCE (INTERNATIONAL) End-to-End Integration of Speech Recognition, Speech Enhancement, and Self-Supervised Learning Representation

Xuankai Chang (Carnegie Mellon University), Takashi Maekaku, Yuya Fujita, Shinji Watanabe (Carnegie Mellon University)

The 23rd Annual Conference of the International Speech Communication Association (INTERSPEECH 2022)

September 19, 2022

This work presents our end-to-end (E2E) automatic speech recognition (ASR) model targetting at robust speech recognition, called Integraded speech Recognition with enhanced speech Input for Self-supervised learning representation (IRIS). Compared with conventional E2E ASR models, the proposed E2E model integrates two important modules including a speech enhancement (SE) module and a self-supervised learning representation (SSLR) module. The SE module enhances the noisy speech. Then the SSLR module extracts features from enhanced speech to be used for speech recognition (ASR). To train the proposed model, we establish an efficient learning scheme. Evaluation results on the monaural CHiME-4 task show that the IRIS model achieves the best performance re- ported in the literature for the single-channel CHiME-4 bench- mark (2.0% for the real development and 3.9% for the real test) thanks to the powerful pre-trained SSLR module and the fine- tuned SE module.

Paper : End-to-End Integration of Speech Recognition, Speech Enhancement, and Self-Supervised Learning Representationopen into new tab or window (external link)