Presentation 2023-03-01
The linguistic influence on speaker verification based on Self-Supervised Learning
Tomoka Wakamatsu, Atsushi Ando, Sayaka Shiota, Ryo Masumura, Hitoshi Kiya,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In recent years, statistical models utilizing Self-Supervised Learning (SSL) have been employed in various fieldsIt has been reported that SSL-based methods achieve high performance in speaker verification. On the other hand, it is known that speaker verification is language-dependent, but the effect of linguistic differences on verification accuracy has not been investigated based on the SSL model. This paper investigates the effect of linguistic differences on accuracy in speaker verification based on the SSL model using English and Japanese speakers in the wild. Experimental results show that the performance of speaker verification based on the SSL model is also degraded by differences in speaker language. We also confirm that the SSL model pre-trained on multilingual dataset reduces the influence of language differences. Moreover, this method records the highest accuracy compared to other conditions when the dataset used for training and evaluation are in the same language.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) speaker verification / Self-Supervised Learning / language dependency / x-vector
Paper # EA2022-118,SIP2022-162,SP2022-82
Date of Issue 2023-02-21 (EA, SIP, SP)

Conference Information
Committee SP / IPSJ-SLP / EA / SIP
Conference Date 2023/2/28(2days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair Tomoki Toda(Nagoya Univ.) / Tomoki Toda(Nagoya Univ.) / Kenichi Furuya(Oita Univ.) / Toshihisa Tanaka(Tokyo Univ. Agri.&Tech.)
Vice Chair / / Tatsuya Kako(NTT) / Junki Ono(Tokyo Metropolitan Univ.) / Koichi Ichige(Yokohama National Univ.) / Takayuki Nakachi(Ryukyu Univ.)
Secretary (NTT) / (Univ. of Electro-Comm.) / Tatsuya Kako(NTT) / Junki Ono(Univ. of Electro-Comm.) / Koichi Ichige(NTT) / Takayuki Nakachi(RitsumeikanUniv.)
Assistant Ryo Aihara(Mitsubishi Electric) / Daisuke Saito(Univ. of Tokyo) / Ryo Aihara(Mitsubishi Electric) / Daisuke Saito(Univ. of Tokyo) / Masato Nakayama(Osaka Sangyo Univ.) / Kouhei Yatabe(Tuat) / Taichi Yoshida(UEC) / Shoko Imaizumi(Chiba Univ.)

Paper Information
Registration To Technical Committee on Speech / Special Interest Group on Spoken Language Processing / Technical Committee on Engineering Acoustics / Technical Committee on Signal Processing
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) The linguistic influence on speaker verification based on Self-Supervised Learning
Sub Title (in English)
Keyword(1) speaker verification
Keyword(2) Self-Supervised Learning
Keyword(3) language dependency
Keyword(4) x-vector
1st Author's Name Tomoka Wakamatsu
1st Author's Affiliation Tokyo Metropolitan University(Tokyo Metropolitan Univ.)
2nd Author's Name Atsushi Ando
2nd Author's Affiliation Nippon Telegraph and Telephone(NTT)
3rd Author's Name Sayaka Shiota
3rd Author's Affiliation Tokyo Metropolitan University(Tokyo Metropolitan Univ.)
4th Author's Name Ryo Masumura
4th Author's Affiliation Nippon Telegraph and Telephone(NTT)
5th Author's Name Hitoshi Kiya
5th Author's Affiliation Tokyo Metropolitan University(Tokyo Metropolitan Univ.)
Date 2023-03-01
Paper # EA2022-118,SIP2022-162,SP2022-82
Volume (vol) vol.122
Number (no) EA-387,SIP-388,SP-389
Page pp.pp.247-252(EA), pp.247-252(SIP), pp.247-252(SP),
#Pages 6
Date of Issue 2023-02-21 (EA, SIP, SP)