Publication Details

Topic identification of spoken documents using unsupervised acoustic unit discovery

KESIRAJU Santosh, PAPPAGARI Raghavendra, ONDEL Yang Lucas Antoine Francois, BURGET Lukáš, DEHAK Najim, KHUDANPUR Sanjeev, ČERNOCKÝ Jan and GANGASHETTY Suryakanth V. Topic identification of spoken documents using unsupervised acoustic unit discovery. In: Proceedings of ICASSP 2017. New Orleans: IEEE Signal Processing Society, 2017, pp. 5745-5749. ISBN 978-1-5090-4117-6.
Czech title
Identifikace témat z mluvených dokumentů pomocí automatického hledání řečových jednotek
Type
conference paper
Language
english
Authors
Kesiraju Santosh (DCGM FIT BUT)
Pappagari Raghavendra (IIIT)
Ondel Yang Lucas Antoine Francois, Mgr., Ph.D. (DCGM FIT BUT)
Burget Lukáš, doc. Ing., Ph.D. (DCGM FIT BUT)
Dehak Najim (JHU)
Khudanpur Sanjeev (JHU)
Černocký Jan, prof. Dr. Ing. (DCGM FIT BUT)
Gangashetty Suryakanth V (IIIT)
URL
Keywords

topic identification, acoustic unit discovery, unsupervised learning, non-parametric Bayesian models

Abstract

This paper investigates the application of unsupervised acoustic unit discovery for topic identification (topic ID) of spoken audio documents. The acoustic unit discovery method is based on a nonparametric Bayesian phone-loop model that segments a speech utterance into phone-like categories. The discovered phone-like (acoustic) units are further fed into the conventional topic ID framework. Using multilingual bottleneck features for the acoustic unit discovery, we show that the proposed method outperforms other systems that are based on cross-lingual phoneme recognizer.

Annotation

This paper investigates the application of unsupervised acoustic unit discovery for topic identification (topic ID) of spoken audio documents. The acoustic unit discovery method is based on a nonparametric Bayesian phone-loop model that segments a speech utterance into phone-like categories. The discovered phone-like (acoustic) units are further fed into the conventional topic ID framework. Using multilingual bottleneck features for the acoustic unit discovery, we show that the proposed method outperforms other systems that are based on cross-lingual phoneme recognizer.

Published
2017
Pages
5745-5749
Proceedings
Proceedings of ICASSP 2017
Conference
2017 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP), New Orleans, USA, US
ISBN
978-1-5090-4117-6
Publisher
IEEE Signal Processing Society
Place
New Orleans, US
DOI
UT WoS
000414286205181
EID Scopus
BibTeX
@INPROCEEDINGS{FITPUB11470,
   author = "Santosh Kesiraju and Raghavendra Pappagari and Francois Antoine Lucas Yang Ondel and Luk\'{a}\v{s} Burget and Najim Dehak and Sanjeev Khudanpur and Jan \v{C}ernock\'{y} and V Suryakanth Gangashetty",
   title = "Topic identification of spoken documents using unsupervised acoustic unit discovery",
   pages = "5745--5749",
   booktitle = "Proceedings of ICASSP 2017",
   year = 2017,
   location = "New Orleans, US",
   publisher = "IEEE Signal Processing Society",
   ISBN = "978-1-5090-4117-6",
   doi = "10.1109/ICASSP.2017.7953257",
   language = "english",
   url = "https://www.fit.vut.cz/research/publication/11470"
}
Back to top