Challenging margin-based speaker embedding extractors by using the variational information bottleneck

Result code in IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216305%3A26230%2F24%3APU154968" target="_blank" >RIV/00216305:26230/24:PU154968 - isvavai.cz</a>

Result on the web

<a href="https://www.isca-archive.org/interspeech_2024/stafylakis24_interspeech.pdf" target="_blank" >https://www.isca-archive.org/interspeech_2024/stafylakis24_interspeech.pdf</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.21437/Interspeech.2024-2058" target="_blank" >10.21437/Interspeech.2024-2058</a>

Result language

angličtina

Original language name

Challenging margin-based speaker embedding extractors by using the variational information bottleneck

Original language description

Speaker embedding extractors are typically trained using a classification loss over the training speakers. During the last few years, the standard softmax/cross-entropy loss has been replaced by the margin-based losses, yielding significant im- provements in speaker recognition accuracy. Motivated by the fact that the margin merely reduces the logit of the target speaker during training, we consider a probabilistic framework that has a similar effect. The variational information bottle- neck provides a principled mechanism for making deterministic nodes stochastic, resulting in an implicit reduction of the pos- terior of the target speaker. We experiment with a wide range of speaker recognition benchmarks and scoring methods and re- port competitive results to those obtained with the state-of-the- art Additive Angular Margin loss.

Czech name

—

Czech description

—

Type

D - Article in proceedings

CEP classification

—

OECD FORD branch

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Project

<a href="/en/project/VB02000060" target="_blank" >VB02000060: Tools To Combat Voice DeepFakes</a><br>

Continuities

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)

Publication year

2024

Confidentiality

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Article name in the collection

Proceedings of Interspeech 2024

ISBN

—

ISSN

1990-9772

e-ISSN

—

Number of pages

5

Pages from-to

3220-3224

Publisher name

International Speech Communication Association

Place of publication

Kos

Event location

Kos

Event date

Sep 1, 2024

Type of event by nationality

WRD - Celosvětová akce

UT code for WoS article

—