Matches in SemOpenAlex for { <https://semopenalex.org/work/W3197227964> ?p ?o ?g. }
Showing items 1 to 48 of
48
with 100 items per page.
- W3197227964 abstract "Training personalized speech enhancement models is innately a no-shot learning problem due to privacy constraints and limited access to noise-free speech from the target user. If there is an abundance of unlabeled noisy speech from the test-time user, a personalized speech enhancement model can be trained using self-supervised learning. One straightforward approach to model personalization is to use the target speaker's noisy recordings as pseudo-sources. Then, a pseudo denoising model learns to remove injected training noises and recover the pseudo-sources. However, this approach is volatile as it depends on the quality of the pseudo-sources, which may be too noisy. As a remedy, we propose an improvement to the self-supervised approach through data purification. We first train an SNR predictor model to estimate the frame-by-frame SNR of the pseudo-sources. Then, the predictor's estimates are converted into weights which adjust the frame-by-frame contribution of the pseudo-sources towards training the personalized model. We empirically show that the proposed data purification step improves the usability of the speaker-specific noisy data in the context of personalized speech enhancement. Without relying on any clean speech recordings or speaker embeddings, our approach may be seen as privacy-preserving." @default.
- W3197227964 created "2021-09-13" @default.
- W3197227964 creator A5064582903 @default.
- W3197227964 creator A5065443361 @default.
- W3197227964 creator A5085011950 @default.
- W3197227964 date "2021-08-30" @default.
- W3197227964 modified "2023-09-25" @default.
- W3197227964 title "Personalized Speech Enhancement Through Self-Supervised Data Augmentation and Purification" @default.
- W3197227964 doi "https://doi.org/10.21437/interspeech.2021-1868" @default.
- W3197227964 hasPublicationYear "2021" @default.
- W3197227964 type Work @default.
- W3197227964 sameAs 3197227964 @default.
- W3197227964 citedByCount "5" @default.
- W3197227964 countsByYear W31972279642021 @default.
- W3197227964 countsByYear W31972279642022 @default.
- W3197227964 crossrefType "proceedings-article" @default.
- W3197227964 hasAuthorship W3197227964A5064582903 @default.
- W3197227964 hasAuthorship W3197227964A5065443361 @default.
- W3197227964 hasAuthorship W3197227964A5085011950 @default.
- W3197227964 hasBestOaLocation W31972279642 @default.
- W3197227964 hasConcept C154945302 @default.
- W3197227964 hasConcept C163294075 @default.
- W3197227964 hasConcept C2776182073 @default.
- W3197227964 hasConcept C28490314 @default.
- W3197227964 hasConcept C41008148 @default.
- W3197227964 hasConceptScore W3197227964C154945302 @default.
- W3197227964 hasConceptScore W3197227964C163294075 @default.
- W3197227964 hasConceptScore W3197227964C2776182073 @default.
- W3197227964 hasConceptScore W3197227964C28490314 @default.
- W3197227964 hasConceptScore W3197227964C41008148 @default.
- W3197227964 hasLocation W31972279641 @default.
- W3197227964 hasLocation W31972279642 @default.
- W3197227964 hasOpenAccess W3197227964 @default.
- W3197227964 hasPrimaryLocation W31972279641 @default.
- W3197227964 hasRelatedWork W1765329493 @default.
- W3197227964 hasRelatedWork W2166831097 @default.
- W3197227964 hasRelatedWork W2312116756 @default.
- W3197227964 hasRelatedWork W2368779261 @default.
- W3197227964 hasRelatedWork W2653598178 @default.
- W3197227964 hasRelatedWork W2778699561 @default.
- W3197227964 hasRelatedWork W2794438528 @default.
- W3197227964 hasRelatedWork W2893763841 @default.
- W3197227964 hasRelatedWork W2995996972 @default.
- W3197227964 hasRelatedWork W3107474891 @default.
- W3197227964 isParatext "false" @default.
- W3197227964 isRetracted "false" @default.
- W3197227964 magId "3197227964" @default.
- W3197227964 workType "article" @default.