Recently, an end-to-end speaker-attributed automatic speech recognition (E2E SA-ASR) model was proposed as a joint model of speaker counting, speech recognition and speaker identification for monaural overlapped speech. In the previous study, the model parameters were trained based on the speaker-attributed maximum mutual information (SA-MMI) criterion, with which the joint posterior probability for multi-talker transcription and speaker identification are maximized over training data. Although SA-MMI training showed promising results for overlapped speech consisting of various numbers of speakers, the training criterion was not directly linked to the final evaluation metric, i.e., speaker-attributed word error rate (SA-WER). In this paper, we propose a speaker-attributed minimum Bayes risk (SA-MBR) training method where the parameters are trained to directly minimize the expected SA-WER over the training data. Experiments using the LibriSpeech corpus show that the proposed SA-MBR training reduces the SA-WER by 9.0 % relative compared with the SA-MMI-trained model.
@article{arxiv.2011.02921,
title = {Minimum Bayes Risk Training for End-to-End Speaker-Attributed ASR},
author = {Naoyuki Kanda and Zhong Meng and Liang Lu and Yashesh Gaur and Xiaofei Wang and Zhuo Chen and Takuya Yoshioka},
journal= {arXiv preprint arXiv:2011.02921},
year = {2020}
}
Comments
Submitted to ICASSP 2021. arXiv admin note: text overlap with arXiv:2006.10930, arXiv:2008.04546