Privacy-oriented manipulation of speaker representations
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2023
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866912021545484288 |
|---|---|
| author | Teixeira, Francisco Abad, Alberto Raj, Bhiksha Trancoso, Isabel |
| author_facet | Teixeira, Francisco Abad, Alberto Raj, Bhiksha Trancoso, Isabel |
| contents | Speaker embeddings are ubiquitous, with applications ranging from speaker recognition and diarization to speech synthesis and voice anonymisation. The amount of information held by these embeddings lends them versatility, but also raises privacy concerns. Speaker embeddings have been shown to contain information on age, sex, health and more, which speakers may want to keep private, especially when this information is not required for the target task. In this work, we propose a method for removing and manipulating private attributes from speaker embeddings that leverages a Vector-Quantized Variational Autoencoder architecture, combined with an adversarial classifier and a novel mutual information loss. We validate our model on two attributes, sex and age, and perform experiments with ignorant and fully-informed attackers, and with in-domain and out-of-domain data. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2310_06652 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | Privacy-oriented manipulation of speaker representations Teixeira, Francisco Abad, Alberto Raj, Bhiksha Trancoso, Isabel Audio and Speech Processing Speaker embeddings are ubiquitous, with applications ranging from speaker recognition and diarization to speech synthesis and voice anonymisation. The amount of information held by these embeddings lends them versatility, but also raises privacy concerns. Speaker embeddings have been shown to contain information on age, sex, health and more, which speakers may want to keep private, especially when this information is not required for the target task. In this work, we propose a method for removing and manipulating private attributes from speaker embeddings that leverages a Vector-Quantized Variational Autoencoder architecture, combined with an adversarial classifier and a novel mutual information loss. We validate our model on two attributes, sex and age, and perform experiments with ignorant and fully-informed attackers, and with in-domain and out-of-domain data. |
| title | Privacy-oriented manipulation of speaker representations |
| topic | Audio and Speech Processing |
| url | https://arxiv.org/abs/2310.06652 |