Data selection by sequence summarizing neural network in mismatch condition training

Kateřina Žmolíková; Martin Karafiát; Karel Veselý; Marc Delcroix; Shinji Watanabe; Lukáš Burget; Jan Cěrnocký

doi:10.21437/Interspeech.2016-741

Data selection by sequence summarizing neural network in mismatch condition training

Kateřina Žmolíková, Martin Karafiát, Karel Veselý, Marc Delcroix, Shinji Watanabe, Lukáš Burget, Jan Cěrnocký

Research output: Contribution to journal › Conference article › peer-review

4 Citations (Scopus)

Abstract

Data augmentation is a simple and efficient technique to improve the robustness of a speech recognizer when deployed in mismatched training-test conditions. Our paper proposes a new approach for selecting data with respect to similarity of acoustic conditions. The similarity is computed based on a sequence summarizing neural network which extracts vectors containing acoustic summary (e.g. noise and reverberation characteristics) of an utterance. Several configurations of this network and different methods of selecting data using these "summary-vectors" were explored. The results are reported on a mismatched condition using AMI training set with the proposed data selection and CHiME3 test set.

Original language	English
Pages (from-to)	2354-2358
Number of pages	5
Journal	Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH
Volume	08-12-September-2016
DOIs	https://doi.org/10.21437/Interspeech.2016-741
Publication status	Published - 2016
Externally published	Yes
Event	17th Annual Conference of the International Speech Communication Association, INTERSPEECH 2016 - San Francisco, United States Duration: 2016 Sept 8 → 2016 Sept 16

Keywords

Automatic speech recognition
Data augmentation
Data selection
Mismatch training condition
Sequence summarization

ASJC Scopus subject areas

Language and Linguistics
Human-Computer Interaction
Signal Processing
Software
Modelling and Simulation

Access to Document

10.21437/Interspeech.2016-741

Cite this

Žmolíková, K., Karafiát, M., Veselý, K., Delcroix, M., Watanabe, S., Burget, L., & Cěrnocký, J. (2016). Data selection by sequence summarizing neural network in mismatch condition training. Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH, 08-12-September-2016, 2354-2358. https://doi.org/10.21437/Interspeech.2016-741

Data selection by sequence summarizing neural network in mismatch condition training. / Žmolíková, Kateřina; Karafiát, Martin; Veselý, Karel et al.

In: Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH, Vol. 08-12-September-2016, 2016, p. 2354-2358.

Research output: Contribution to journal › Conference article › peer-review

Žmolíková, K, Karafiát, M, Veselý, K, Delcroix, M, Watanabe, S, Burget, L & Cěrnocký, J 2016, 'Data selection by sequence summarizing neural network in mismatch condition training', Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH, vol. 08-12-September-2016, pp. 2354-2358. https://doi.org/10.21437/Interspeech.2016-741

@article{6ddba5c309bb4ee1817487b6c7f62a20,

title = "Data selection by sequence summarizing neural network in mismatch condition training",

abstract = "Data augmentation is a simple and efficient technique to improve the robustness of a speech recognizer when deployed in mismatched training-test conditions. Our paper proposes a new approach for selecting data with respect to similarity of acoustic conditions. The similarity is computed based on a sequence summarizing neural network which extracts vectors containing acoustic summary (e.g. noise and reverberation characteristics) of an utterance. Several configurations of this network and different methods of selecting data using these {"}summary-vectors{"} were explored. The results are reported on a mismatched condition using AMI training set with the proposed data selection and CHiME3 test set.",

keywords = "Automatic speech recognition, Data augmentation, Data selection, Mismatch training condition, Sequence summarization",

author = "Kate{\v r}ina {\v Z}mol{\'i}kov{\'a} and Martin Karafi{\'a}t and Karel Vesel{\'y} and Marc Delcroix and Shinji Watanabe and Luk{\'a}{\v s} Burget and Jan C{\v e}rnock{\'y}",

note = "Publisher Copyright: Copyright {\textcopyright} 2016 ISCA.; 17th Annual Conference of the International Speech Communication Association, INTERSPEECH 2016 ; Conference date: 08-09-2016 Through 16-09-2016",

year = "2016",

doi = "10.21437/Interspeech.2016-741",

language = "English",

volume = "08-12-September-2016",

pages = "2354--2358",

journal = "Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH",

issn = "2308-457X",

}

TY - JOUR

T1 - Data selection by sequence summarizing neural network in mismatch condition training

AU - Žmolíková, Kateřina

AU - Karafiát, Martin

AU - Veselý, Karel

AU - Delcroix, Marc

AU - Watanabe, Shinji

AU - Burget, Lukáš

AU - Cěrnocký, Jan

PY - 2016

Y1 - 2016

N2 - Data augmentation is a simple and efficient technique to improve the robustness of a speech recognizer when deployed in mismatched training-test conditions. Our paper proposes a new approach for selecting data with respect to similarity of acoustic conditions. The similarity is computed based on a sequence summarizing neural network which extracts vectors containing acoustic summary (e.g. noise and reverberation characteristics) of an utterance. Several configurations of this network and different methods of selecting data using these "summary-vectors" were explored. The results are reported on a mismatched condition using AMI training set with the proposed data selection and CHiME3 test set.

AB - Data augmentation is a simple and efficient technique to improve the robustness of a speech recognizer when deployed in mismatched training-test conditions. Our paper proposes a new approach for selecting data with respect to similarity of acoustic conditions. The similarity is computed based on a sequence summarizing neural network which extracts vectors containing acoustic summary (e.g. noise and reverberation characteristics) of an utterance. Several configurations of this network and different methods of selecting data using these "summary-vectors" were explored. The results are reported on a mismatched condition using AMI training set with the proposed data selection and CHiME3 test set.

KW - Automatic speech recognition

KW - Data augmentation

KW - Data selection

KW - Mismatch training condition

KW - Sequence summarization

UR - http://www.scopus.com/inward/record.url?scp=84994382229&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84994382229&partnerID=8YFLogxK

U2 - 10.21437/Interspeech.2016-741

DO - 10.21437/Interspeech.2016-741

M3 - Conference article

AN - SCOPUS:84994382229

SN - 2308-457X

VL - 08-12-September-2016

SP - 2354

EP - 2358

JO - Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH

JF - Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH

T2 - 17th Annual Conference of the International Speech Communication Association, INTERSPEECH 2016

Y2 - 8 September 2016 through 16 September 2016

ER -

Data selection by sequence summarizing neural network in mismatch condition training

Abstract

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this