Matches in SemOpenAlex for { <https://semopenalex.org/work/W2787447541> ?p ?o ?g. }
Showing items 1 to 85 of
85
with 100 items per page.
- W2787447541 abstract "This paper describes our unsupervised subword modeling pipeline for the zero resource speech challenge (ZeroSpeech) 2017. Our approach is built around the Dirichlet process Gaussian mixture model (DPGMM) that we use to cluster speech feature vectors into a dynamically sized set of classes. By considering each class an acoustic unit, speech can be represented as sequence of class posteriorgrams. We enhance this method by automatically optimizing the DPGMM sampler's input features in a multi-stage clustering framework, where we unsupervisedly learn transformations using LDA, MLLT and (basis) fMLLR to reduce variance in the features. We show that this optimization considerably boosts the subword modeling quality, according to the performance on the ABX phone discriminability task. For the first time, we apply inferred subword models to previously unseen data from a new set of speakers. We demonstrate our method's good generalization and the effectiveness of its blind speaker adaptation in extensive experiments on a multitude of datasets. Our pipeline has very little need for hyper-parameter adjustment and is entirely unsupervised, i.e., it only takes raw audio recordings as input, without requiring any pre-defined segmentation, explicit speaker IDs or other meta data." @default.
- W2787447541 created "2018-02-23" @default.
- W2787447541 creator A5018909877 @default.
- W2787447541 creator A5020994673 @default.
- W2787447541 creator A5040108974 @default.
- W2787447541 date "2017-12-01" @default.
- W2787447541 modified "2023-09-29" @default.
- W2787447541 title "Feature optimized DPGMM clustering for unsupervised subword modeling: A contribution to zerospeech 2017" @default.
- W2787447541 cites W1599512239 @default.
- W2787447541 cites W1631260214 @default.
- W2787447541 cites W2001619934 @default.
- W2787447541 cites W2002342963 @default.
- W2787447541 cites W2019042707 @default.
- W2787447541 cites W2078769636 @default.
- W2787447541 cites W2106554350 @default.
- W2787447541 cites W2113641473 @default.
- W2787447541 cites W2124629003 @default.
- W2787447541 cites W2509930204 @default.
- W2787447541 cites W2963620343 @default.
- W2787447541 cites W2586754519 @default.
- W2787447541 doi "https://doi.org/10.1109/asru.2017.8269011" @default.
- W2787447541 hasPublicationYear "2017" @default.
- W2787447541 type Work @default.
- W2787447541 sameAs 2787447541 @default.
- W2787447541 citedByCount "39" @default.
- W2787447541 countsByYear W27874475412017 @default.
- W2787447541 countsByYear W27874475412018 @default.
- W2787447541 countsByYear W27874475412019 @default.
- W2787447541 countsByYear W27874475412020 @default.
- W2787447541 countsByYear W27874475412021 @default.
- W2787447541 countsByYear W27874475412022 @default.
- W2787447541 countsByYear W27874475412023 @default.
- W2787447541 crossrefType "proceedings-article" @default.
- W2787447541 hasAuthorship W2787447541A5018909877 @default.
- W2787447541 hasAuthorship W2787447541A5020994673 @default.
- W2787447541 hasAuthorship W2787447541A5040108974 @default.
- W2787447541 hasConcept C138885662 @default.
- W2787447541 hasConcept C151989614 @default.
- W2787447541 hasConcept C153180895 @default.
- W2787447541 hasConcept C154945302 @default.
- W2787447541 hasConcept C199360897 @default.
- W2787447541 hasConcept C23224414 @default.
- W2787447541 hasConcept C2776401178 @default.
- W2787447541 hasConcept C2778724510 @default.
- W2787447541 hasConcept C28490314 @default.
- W2787447541 hasConcept C41008148 @default.
- W2787447541 hasConcept C41895202 @default.
- W2787447541 hasConcept C43521106 @default.
- W2787447541 hasConcept C52622490 @default.
- W2787447541 hasConcept C61224824 @default.
- W2787447541 hasConcept C73555534 @default.
- W2787447541 hasConcept C8038995 @default.
- W2787447541 hasConceptScore W2787447541C138885662 @default.
- W2787447541 hasConceptScore W2787447541C151989614 @default.
- W2787447541 hasConceptScore W2787447541C153180895 @default.
- W2787447541 hasConceptScore W2787447541C154945302 @default.
- W2787447541 hasConceptScore W2787447541C199360897 @default.
- W2787447541 hasConceptScore W2787447541C23224414 @default.
- W2787447541 hasConceptScore W2787447541C2776401178 @default.
- W2787447541 hasConceptScore W2787447541C2778724510 @default.
- W2787447541 hasConceptScore W2787447541C28490314 @default.
- W2787447541 hasConceptScore W2787447541C41008148 @default.
- W2787447541 hasConceptScore W2787447541C41895202 @default.
- W2787447541 hasConceptScore W2787447541C43521106 @default.
- W2787447541 hasConceptScore W2787447541C52622490 @default.
- W2787447541 hasConceptScore W2787447541C61224824 @default.
- W2787447541 hasConceptScore W2787447541C73555534 @default.
- W2787447541 hasConceptScore W2787447541C8038995 @default.
- W2787447541 hasLocation W27874475411 @default.
- W2787447541 hasOpenAccess W2787447541 @default.
- W2787447541 hasPrimaryLocation W27874475411 @default.
- W2787447541 hasRelatedWork W1512197020 @default.
- W2787447541 hasRelatedWork W1771058801 @default.
- W2787447541 hasRelatedWork W1981297204 @default.
- W2787447541 hasRelatedWork W2014114590 @default.
- W2787447541 hasRelatedWork W2098258572 @default.
- W2787447541 hasRelatedWork W2162966498 @default.
- W2787447541 hasRelatedWork W2165240440 @default.
- W2787447541 hasRelatedWork W2536059291 @default.
- W2787447541 hasRelatedWork W2548170873 @default.
- W2787447541 hasRelatedWork W3106913486 @default.
- W2787447541 isParatext "false" @default.
- W2787447541 isRetracted "false" @default.
- W2787447541 magId "2787447541" @default.
- W2787447541 workType "article" @default.