Acoustic model adaptation from raw waveforms with Sincnet

Joachim Fainberg; Ondrej Klejch; Erfan Loweimi; Peter Bell; Steve Renals

doi:10.1109/ASRU46091.2019.9003974

Acoustic model adaptation from raw waveforms with Sincnet

Joachim Fainberg, Ondrej Klejch, Erfan Loweimi, Peter Bell, Steve Renals

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract / Description of output

Raw waveform acoustic modelling has recently gained interest due to neural networks’ ability to learn feature extraction, and the potential for finding better representations for a given scenario than hand-crafted features. SincNet has been proposed to reduce the number of parameters required in rawwaveform modelling, by restricting the filter functions, rather than having to learn every tap of each filter. We study the adaptation of the SincNet filter parameters from adults’ to children’s speech, and show that the parameterisation of the SincNet layer is well suited for adaptation in practice: we can efficiently adapt with a very small number of parameters, producing error rates comparable to techniques using orders of magnitude more parameters.

Original language	English
Title of host publication	2019 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU)
Publisher	Institute of Electrical and Electronics Engineers
Pages	897-904
Number of pages	8
ISBN (Electronic)	978-1-7281-0306-8
ISBN (Print)	978-1-7281-0307-5
DOIs	https://doi.org/10.1109/ASRU46091.2019.9003974
Publication status	Published - 20 Feb 2020
Event	IEEE Automatic Speech Recognition and Understanding Workshop 2019 - Sentosa, Singapore Duration: 14 Dec 2019 → 18 Dec 2019 http://asru2019.org/wp/

Conference

Conference	IEEE Automatic Speech Recognition and Understanding Workshop 2019
Abbreviated title	ASRU 2019
Country/Territory	Singapore
City	Sentosa
Period	14/12/19 → 18/12/19
Internet address	http://asru2019.org/wp/

Keywords / Materials (for Non-textual outputs)

Acoustic model adaptation
children’s speech
raw waveform
SincNet

Access to Document

10.1109/ASRU46091.2019.9003974Licence: All Rights Reserved

Acoustic Model Adaptation_FAINBERG_DOA13092019_AFVAccepted author manuscript, 285 KB

https://ieeexplore.ieee.org/document/9003974Licence: All Rights Reserved

ELG- European Language Grid
Germann, U. & Renals, S.
EU government bodies
1/01/19 → 30/06/22
Project: Research
SpeechWave
Renals, S. & Bell, P.
EPSRC
1/03/18 → 21/05/22
Project: Research
Multi-domain speech recognition
Renals, S.
Non-EU industry, commerce and public corporations
1/09/15 → 28/02/19
Project: Research

Cite this

@inproceedings{b41886f687a645a5a9e4ad51eec508c0,

title = "Acoustic model adaptation from raw waveforms with Sincnet",

abstract = "Raw waveform acoustic modelling has recently gained interest due to neural networks{\textquoteright} ability to learn feature extraction, and the potential for finding better representations for a given scenario than hand-crafted features. SincNet has been proposed to reduce the number of parameters required in rawwaveform modelling, by restricting the filter functions, rather than having to learn every tap of each filter. We study the adaptation of the SincNet filter parameters from adults{\textquoteright} to children{\textquoteright}s speech, and show that the parameterisation of the SincNet layer is well suited for adaptation in practice: we can efficiently adapt with a very small number of parameters, producing error rates comparable to techniques using orders of magnitude more parameters.",

keywords = "Acoustic model adaptation, children{\textquoteright}s speech, raw waveform, SincNet",

author = "Joachim Fainberg and Ondrej Klejch and Erfan Loweimi and Peter Bell and Steve Renals",

year = "2020",

month = feb,

day = "20",

doi = "10.1109/ASRU46091.2019.9003974",

language = "English",

isbn = "978-1-7281-0307-5",

pages = "897--904",

booktitle = "2019 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU)",

publisher = "Institute of Electrical and Electronics Engineers",

note = "IEEE Automatic Speech Recognition and Understanding Workshop 2019, ASRU 2019 ; Conference date: 14-12-2019 Through 18-12-2019",

url = "http://asru2019.org/wp/",

}

Fainberg, J, Klejch, O , Loweimi, E , Bell, P & Renals, S 2020, Acoustic model adaptation from raw waveforms with Sincnet. in 2019 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU). Institute of Electrical and Electronics Engineers, pp. 897-904, IEEE Automatic Speech Recognition and Understanding Workshop 2019, Sentosa, Singapore, 14/12/19. https://doi.org/10.1109/ASRU46091.2019.9003974

TY - GEN

T1 - Acoustic model adaptation from raw waveforms with Sincnet

AU - Fainberg, Joachim

AU - Klejch, Ondrej

AU - Loweimi, Erfan

AU - Bell, Peter

AU - Renals, Steve

PY - 2020/2/20

Y1 - 2020/2/20

N2 - Raw waveform acoustic modelling has recently gained interest due to neural networks’ ability to learn feature extraction, and the potential for finding better representations for a given scenario than hand-crafted features. SincNet has been proposed to reduce the number of parameters required in rawwaveform modelling, by restricting the filter functions, rather than having to learn every tap of each filter. We study the adaptation of the SincNet filter parameters from adults’ to children’s speech, and show that the parameterisation of the SincNet layer is well suited for adaptation in practice: we can efficiently adapt with a very small number of parameters, producing error rates comparable to techniques using orders of magnitude more parameters.

AB - Raw waveform acoustic modelling has recently gained interest due to neural networks’ ability to learn feature extraction, and the potential for finding better representations for a given scenario than hand-crafted features. SincNet has been proposed to reduce the number of parameters required in rawwaveform modelling, by restricting the filter functions, rather than having to learn every tap of each filter. We study the adaptation of the SincNet filter parameters from adults’ to children’s speech, and show that the parameterisation of the SincNet layer is well suited for adaptation in practice: we can efficiently adapt with a very small number of parameters, producing error rates comparable to techniques using orders of magnitude more parameters.

KW - Acoustic model adaptation

KW - children’s speech

KW - raw waveform

KW - SincNet

U2 - 10.1109/ASRU46091.2019.9003974

DO - 10.1109/ASRU46091.2019.9003974

M3 - Conference contribution

SN - 978-1-7281-0307-5

SP - 897

EP - 904

BT - 2019 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU)

PB - Institute of Electrical and Electronics Engineers

T2 - IEEE Automatic Speech Recognition and Understanding Workshop 2019

Y2 - 14 December 2019 through 18 December 2019

ER -

Acoustic model adaptation from raw waveforms with Sincnet

Abstract / Description of output

Conference

Keywords / Materials (for Non-textual outputs)

Access to Document

Fingerprint

Projects

ELG- European Language Grid

SpeechWave

Multi-domain speech recognition

Cite this