Publication Details

An Empirical evaluation of zero resource acoustic unit discovery

LIU, C.; YANG, J.; SUN, M.; KESIRAJU, S.; ROTT, A.; ONDEL YANG, L.; GHAHREMANI, P.; DEHAK, N.; BURGET, L.; KHUDANPUR, S. An Empirical evaluation of zero resource acoustic unit discovery. In Proceedings of ICASSP 2017. New Orleans: IEEE Signal Processing Society, 2017. p. 5305-5309. ISBN: 978-1-5090-4117-6.
Czech title
Empirické hodnocení automatického hledání řečových jednotek bez popsaných trénovacích dat
Type
conference paper
Language
English
Authors
URL
Keywords

Acoustic unit discovery, unsupervised linear discriminant analysis, evaluation methods, zero resource

Abstract

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.

Annotation

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.

Published
2017
Pages
5305–5309
Proceedings
Proceedings of ICASSP 2017
Conference
42nd IEEE International Conference on Acoustics, Speech and Signal Processing, New Orleans, USA, US
ISBN
978-1-5090-4117-6
Publisher
IEEE Signal Processing Society
Place
New Orleans
DOI
UT WoS
000414286205093
EID Scopus
BibTeX
@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/"
}
Back to top