# End-to-End Integration of Speech Recognition, Speech Enhancement, and Self-Supervised Learning Representation

Source: [End-to-End Integration of Speech Recognition, Speech Enhancement, and Self-Supervised Learning Representation](<https://www.sota2.com/research/paper/68fa6c311e209131cea32bf7>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-28T13:56:52.681Z

Authors: Xuankai Chang, Takashi Maekaku, Yuya Fujita, Shinji Watanabe

Published: 2022-04-01T16:02:31.000Z

## Abstract

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 reported in the literature for the single-channel CHiME-4 benchmark (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.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Automatic Speech Recognition on CHiME-4 single-channel (dt\_real)](<https://www.sota2.com/research/sota/automatic-speech-recognition-on-chime-4-single-channel-dt_real.md>): Automatic Speech Recognition; CHiME-4 single-channel (dt\_real); WER: 2.

- [Automatic Speech Recognition on CHiME-4 single-channel (et\_simu)](<https://www.sota2.com/research/sota/automatic-speech-recognition-on-chime-4-single-channel-et_simu.md>): Automatic Speech Recognition; CHiME-4 single-channel (et\_simu); WER: 0.061.

- [Automatic Speech Recognition on CHiME-4 single-channel (dt\_simu)](<https://www.sota2.com/research/sota/automatic-speech-recognition-on-chime-4-single-channel-dt_simu.md>): Automatic Speech Recognition; CHiME-4 single-channel (dt\_simu); WER: 3.2.

- [Automatic Speech Recognition on CHiME-4 single-channel (et\_real)](<https://www.sota2.com/research/sota/automatic-speech-recognition-on-chime-4-single-channel-et_real.md>): Automatic Speech Recognition; CHiME-4 single-channel (et\_real); WER: 3.9.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2204.00540>)

[PDF](<https://arxiv.org/pdf/2204.00540.pdf>)