Matches in SemOpenAlex for { <https://semopenalex.org/work/W4226185896> ?p ?o ?g. }
Showing items 1 to 47 of
47
with 100 items per page.
- W4226185896 abstract "This paper investigates how to improve the runtime speed of personalized speech enhancement (PSE) networks while maintaining the model quality. Our approach includes two aspects: architecture and knowledge distillation (KD). We propose an end-to-end enhancement (E3Net) model architecture, which is $3times$ faster than a baseline STFT-based model. Besides, we use KD techniques to develop compressed student models without significantly degrading quality. In addition, we investigate using noisy data without reference clean signals for training the student models, where we combine KD with multi-task learning (MTL) using automatic speech recognition (ASR) loss. Our results show that E3Net provides better speech and transcription quality with a lower target speaker over-suppression (TSOS) rate than the baseline model. Furthermore, we show that the KD methods can yield student models that are $2-4times$ faster than the teacher and provides reasonable quality. Combining KD and MTL improves the ASR and TSOS metrics without degrading the speech quality." @default.
- W4226185896 created "2022-05-05" @default.
- W4226185896 creator A5019074126 @default.
- W4226185896 creator A5026088950 @default.
- W4226185896 creator A5028363114 @default.
- W4226185896 creator A5032260110 @default.
- W4226185896 date "2022-09-18" @default.
- W4226185896 modified "2023-10-16" @default.
- W4226185896 title "Fast Real-time Personalized Speech Enhancement: End-to-End Enhancement Network (E3Net) and Knowledge Distillation" @default.
- W4226185896 doi "https://doi.org/10.21437/interspeech.2022-10962" @default.
- W4226185896 hasPublicationYear "2022" @default.
- W4226185896 type Work @default.
- W4226185896 citedByCount "6" @default.
- W4226185896 countsByYear W42261858962023 @default.
- W4226185896 crossrefType "proceedings-article" @default.
- W4226185896 hasAuthorship W4226185896A5019074126 @default.
- W4226185896 hasAuthorship W4226185896A5026088950 @default.
- W4226185896 hasAuthorship W4226185896A5028363114 @default.
- W4226185896 hasAuthorship W4226185896A5032260110 @default.
- W4226185896 hasBestOaLocation W42261858962 @default.
- W4226185896 hasConcept C154945302 @default.
- W4226185896 hasConcept C163294075 @default.
- W4226185896 hasConcept C2776182073 @default.
- W4226185896 hasConcept C41008148 @default.
- W4226185896 hasConcept C74296488 @default.
- W4226185896 hasConceptScore W4226185896C154945302 @default.
- W4226185896 hasConceptScore W4226185896C163294075 @default.
- W4226185896 hasConceptScore W4226185896C2776182073 @default.
- W4226185896 hasConceptScore W4226185896C41008148 @default.
- W4226185896 hasConceptScore W4226185896C74296488 @default.
- W4226185896 hasLocation W42261858961 @default.
- W4226185896 hasLocation W42261858962 @default.
- W4226185896 hasOpenAccess W4226185896 @default.
- W4226185896 hasPrimaryLocation W42261858961 @default.
- W4226185896 hasRelatedWork W1549018748 @default.
- W4226185896 hasRelatedWork W2001712873 @default.
- W4226185896 hasRelatedWork W2034282843 @default.
- W4226185896 hasRelatedWork W2043060026 @default.
- W4226185896 hasRelatedWork W2119252293 @default.
- W4226185896 hasRelatedWork W2312526371 @default.
- W4226185896 hasRelatedWork W2774176625 @default.
- W4226185896 hasRelatedWork W3015352480 @default.
- W4226185896 hasRelatedWork W4226185896 @default.
- W4226185896 hasRelatedWork W4296123605 @default.
- W4226185896 isParatext "false" @default.
- W4226185896 isRetracted "false" @default.
- W4226185896 workType "article" @default.