Publication result detail

Training Speaker Embedding Extractors Using Multi-Speaker Audio with Unknown Speaker Boundaries

STAFYLAKIS, T.; MOŠNER, L.; PLCHOT, O.; ROHDIN, J.; SILNOVA, A.; BURGET, L.; ČERNOCKÝ, J.

Original Title

Training Speaker Embedding Extractors Using Multi-Speaker Audio with Unknown Speaker Boundaries

English Title

Training Speaker Embedding Extractors Using Multi-Speaker Audio with Unknown Speaker Boundaries

Type

Paper in proceedings (conference paper)

Original Abstract

In this paper, we demonstrate a method for training speaker em- bedding extractors using weak annotation. More specifically, we are using the full VoxCeleb recordings and the name of the celebrities appearing on each video without knowledge of the time intervals the celebrities appear in the video. We show that by combining a baseline speaker diarization algorithm that re- quires no training or parameter tuning, a modified loss with aggregation over segments, and a two-stage training approach, we are able to train a competitive ResNet-based embedding extractor. Finally, we experiment with two different aggregation functions and analyze their behaviour in terms of their gradients.

English abstract

In this paper, we demonstrate a method for training speaker em- bedding extractors using weak annotation. More specifically, we are using the full VoxCeleb recordings and the name of the celebrities appearing on each video without knowledge of the time intervals the celebrities appear in the video. We show that by combining a baseline speaker diarization algorithm that re- quires no training or parameter tuning, a modified loss with aggregation over segments, and a two-stage training approach, we are able to train a competitive ResNet-based embedding extractor. Finally, we experiment with two different aggregation functions and analyze their behaviour in terms of their gradients.

Keywords

Speaker Embedding Extractors, Multi-Speaker Audio, Unknown Speaker Boundaries

Key words in English

Speaker Embedding Extractors, Multi-Speaker Audio, Unknown Speaker Boundaries

Authors

STAFYLAKIS, T.; MOŠNER, L.; PLCHOT, O.; ROHDIN, J.; SILNOVA, A.; BURGET, L.; ČERNOCKÝ, J.

RIV year

2023

Released

18.09.2022

Publisher

International Speech Communication Association

Location

Incheon

Book

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

ISBN

1990-9772

Periodical

Proceedings of Interspeech

Volume

2022

Number

9

State

French Republic

Pages from

605

Pages to

609

Pages count

5

URL

https://www.isca-speech.org/archive/pdfs/interspeech_2022/stafylakis22_interspeech.pdf

BibTex

@inproceedings{BUT179781,
  author="Themos {Stafylakis} and Ladislav {Mošner} and Oldřich {Plchot} and Johan Andréas {Rohdin} and Anna {Silnova} and Lukáš {Burget} and Jan {Černocký}",
  title="Training Speaker Embedding Extractors Using Multi-Speaker Audio with Unknown Speaker Boundaries",
  booktitle="Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH",
  year="2022",
  journal="Proceedings of Interspeech",
  volume="2022",
  number="9",
  pages="605--609",
  publisher="International Speech Communication Association",
  address="Incheon",
  doi="10.21437/Interspeech.2022-10165",
  issn="1990-9772",
  url="https://www.isca-speech.org/archive/pdfs/interspeech_2022/stafylakis22_interspeech.pdf"
}

Documents

stafylakis_interspeech2022_training