Matches in SemOpenAlex for { <https://semopenalex.org/work/W3127172085> ?p ?o ?g. }
Showing items 1 to 95 of
95
with 100 items per page.
- W3127172085 abstract "We study permutation invariant training (PIT), which targets at the permutation ambiguity problem for speaker independent source separation models. We extend two state-of-the-art PIT strategies. First, we look at the two-stage speaker separation and tracking algorithm based on frame level PIT (tPIT) and clustering, which was originally proposed for the STFT domain, and we adapt it to work with waveforms and over a learned latent space. Further, we propose an efficient clustering loss scalable to waveform models. Second, we extend a recently proposed auxiliary speaker-ID loss with a deep feature loss based on problem agnostic speech features, to reduce the local permutation errors made by the utterance level PIT (uPIT). Our results show that the proposed extensions help reducing permutation ambiguity. However, we also note that the studied STFT-based models are more effective at reducing permutation errors than waveform-based models, a perspective overlooked in recent studies." @default.
- W3127172085 created "2021-02-15" @default.
- W3127172085 creator A5068529504 @default.
- W3127172085 creator A5086418698 @default.
- W3127172085 date "2021-02-09" @default.
- W3127172085 modified "2023-09-27" @default.
- W3127172085 title "On permutation invariant training for speech source separation" @default.
- W3127172085 cites W2221409856 @default.
- W3127172085 cites W2734774145 @default.
- W3127172085 cites W2800022361 @default.
- W3127172085 cites W2890111732 @default.
- W3127172085 cites W2944972166 @default.
- W3127172085 cites W2952218014 @default.
- W3127172085 cites W2962788625 @default.
- W3127172085 cites W2962935966 @default.
- W3127172085 cites W2963921132 @default.
- W3127172085 cites W2964058413 @default.
- W3127172085 cites W2972382840 @default.
- W3127172085 cites W2972460025 @default.
- W3127172085 cites W2972693890 @default.
- W3127172085 cites W2972864514 @default.
- W3127172085 cites W2973157397 @default.
- W3127172085 cites W2981436548 @default.
- W3127172085 cites W2981976899 @default.
- W3127172085 cites W2990666817 @default.
- W3127172085 cites W2998657200 @default.
- W3127172085 cites W3004940340 @default.
- W3127172085 cites W3015213852 @default.
- W3127172085 cites W3015225820 @default.
- W3127172085 cites W3015605047 @default.
- W3127172085 cites W3015843733 @default.
- W3127172085 cites W3027008958 @default.
- W3127172085 cites W3035268204 @default.
- W3127172085 doi "https://doi.org/10.48550/arxiv.2102.04945" @default.
- W3127172085 hasPublicationYear "2021" @default.
- W3127172085 type Work @default.
- W3127172085 sameAs 3127172085 @default.
- W3127172085 citedByCount "0" @default.
- W3127172085 crossrefType "posted-content" @default.
- W3127172085 hasAuthorship W3127172085A5068529504 @default.
- W3127172085 hasAuthorship W3127172085A5086418698 @default.
- W3127172085 hasBestOaLocation W31271720851 @default.
- W3127172085 hasConcept C11413529 @default.
- W3127172085 hasConcept C121332964 @default.
- W3127172085 hasConcept C138885662 @default.
- W3127172085 hasConcept C153180895 @default.
- W3127172085 hasConcept C154945302 @default.
- W3127172085 hasConcept C190470478 @default.
- W3127172085 hasConcept C199360897 @default.
- W3127172085 hasConcept C21308566 @default.
- W3127172085 hasConcept C24890656 @default.
- W3127172085 hasConcept C2775852435 @default.
- W3127172085 hasConcept C2776401178 @default.
- W3127172085 hasConcept C2780522230 @default.
- W3127172085 hasConcept C28490314 @default.
- W3127172085 hasConcept C33923547 @default.
- W3127172085 hasConcept C37914503 @default.
- W3127172085 hasConcept C41008148 @default.
- W3127172085 hasConcept C41895202 @default.
- W3127172085 hasConcept C73555534 @default.
- W3127172085 hasConceptScore W3127172085C11413529 @default.
- W3127172085 hasConceptScore W3127172085C121332964 @default.
- W3127172085 hasConceptScore W3127172085C138885662 @default.
- W3127172085 hasConceptScore W3127172085C153180895 @default.
- W3127172085 hasConceptScore W3127172085C154945302 @default.
- W3127172085 hasConceptScore W3127172085C190470478 @default.
- W3127172085 hasConceptScore W3127172085C199360897 @default.
- W3127172085 hasConceptScore W3127172085C21308566 @default.
- W3127172085 hasConceptScore W3127172085C24890656 @default.
- W3127172085 hasConceptScore W3127172085C2775852435 @default.
- W3127172085 hasConceptScore W3127172085C2776401178 @default.
- W3127172085 hasConceptScore W3127172085C2780522230 @default.
- W3127172085 hasConceptScore W3127172085C28490314 @default.
- W3127172085 hasConceptScore W3127172085C33923547 @default.
- W3127172085 hasConceptScore W3127172085C37914503 @default.
- W3127172085 hasConceptScore W3127172085C41008148 @default.
- W3127172085 hasConceptScore W3127172085C41895202 @default.
- W3127172085 hasConceptScore W3127172085C73555534 @default.
- W3127172085 hasLocation W31271720851 @default.
- W3127172085 hasOpenAccess W3127172085 @default.
- W3127172085 hasPrimaryLocation W31271720851 @default.
- W3127172085 hasRelatedWork W1972370106 @default.
- W3127172085 hasRelatedWork W2015538044 @default.
- W3127172085 hasRelatedWork W2025991752 @default.
- W3127172085 hasRelatedWork W2052253960 @default.
- W3127172085 hasRelatedWork W2171511892 @default.
- W3127172085 hasRelatedWork W2319337512 @default.
- W3127172085 hasRelatedWork W2382607599 @default.
- W3127172085 hasRelatedWork W2970216048 @default.
- W3127172085 hasRelatedWork W62672326 @default.
- W3127172085 hasRelatedWork W2480412556 @default.
- W3127172085 isParatext "false" @default.
- W3127172085 isRetracted "false" @default.
- W3127172085 magId "3127172085" @default.
- W3127172085 workType "article" @default.