Publication Details
An Empirical evaluation of zero resource acoustic unit discovery
Yang Jinyi (FIT)
Sun Ming (FIT)
Kesiraju Santosh, Ph.D. (DCGM)
Rott Alena (FIT)
Ondel Lucas Antoine Francois, Mgr., Ph.D. (SSDIT)
Ghahremani Pegah (FIT)
Dehak Najim
Burget Lukáš, doc. Ing., Ph.D. (DCGM)
Khudanpur Sanjeev
Acoustic unit discovery, unsupervised linear discriminant analysis, evaluation methods, zero resource
Acoustic unit discovery (AUD) is a process of automatically identifying a categorical acoustic unit inventory from speech and producing corresponding acoustic unit tokenizations. AUD provides an important avenue for unsupervised acoustic model training in a zero resource setting where expert-provided linguistic knowledge and transcribed speech are unavailable. Therefore, to further facilitate zero-resource AUD process, in this paper, we demonstrate acoustic feature representations can be significantly improved by (i) performing linear discriminant analysis (LDA) in an unsupervised self-trained fashion, and (ii) leveraging resources of other languages through building a multilingual bottleneck (BN) feature extractor to give effective cross-lingual generalization. Moreover, we perform comprehensive evaluations of AUD efficacy on multiple downstream speech applications, and their correlated performance suggests that AUD evaluations are feasible using different alternative language resources when only a subset of these evaluation resources can be available in typical zero resource applications.
Acoustic unit discovery (AUD) is a process of automatically identifying a categorical acoustic unit inventory from speech and producing corresponding acoustic unit tokenizations. AUD provides an important avenue for unsupervised acoustic model training in a zero resource setting where expert-provided linguistic knowledge and transcribed speech are unavailable. Therefore, to further facilitate zero-resource AUD process, in this paper, we demonstrate acoustic feature representations can be significantly improved by (i) performing linear discriminant analysis (LDA) in an unsupervised self-trained fashion, and (ii) leveraging resources of other languages through building a multilingual bottleneck (BN) feature extractor to give effective cross-lingual generalization. Moreover, we perform comprehensive evaluations of AUD efficacy on multiple downstream speech applications, and their correlated performance suggests that AUD evaluations are feasible using different alternative language resources when only a subset of these evaluation resources can be available in typical zero resource applications.
@inproceedings{BUT144451,
author="Chunxi {Liu} and Jinyi {Yang} and Ming {Sun} and Santosh {Kesiraju} and Alena {Rott} and Lucas Antoine Francois {Ondel} and Pegah {Ghahremani} and Najim {Dehak} and Lukáš {Burget} and Sanjeev {Khudanpur}",
title="An Empirical evaluation of zero resource acoustic unit discovery",
booktitle="Proceedings of ICASSP 2017",
year="2017",
pages="5305--5309",
publisher="IEEE Signal Processing Society",
address="New Orleans",
doi="10.1109/ICASSP.2017.7953169",
isbn="978-1-5090-4117-6",
url="https://www.fit.vut.cz/research/publication/11471/"
}