Matches in SemOpenAlex for { <https://semopenalex.org/work/W3110761489> ?p ?o ?g. }
- W3110761489 abstract "We propose a new unsupervised model for mapping a variable-duration speech segment to a fixed-dimensional representation. The resulting acoustic word embeddings can form the basis of search, discovery, and indexing systems for low- and zero-resource languages. Our model, which we refer to as a maximal sampling correspondence variational autoencoder (MCVAE), is a recurrent neural network (RNN) trained with a novel self-supervised correspondence loss that encourages consistency between embeddings of different instances of the same word. Our training scheme improves on previous correspondence training approaches through the use and comparison of multiple samples from the approximate posterior distribution. In the zero-resource setting, the MCVAE can be trained in an unsupervised way, without any ground-truth word pairs, by using the word-like segments discovered via an unsupervised term discovery system. In both this setting and a semi-supervised low-resource setting (with a limited set of ground-truth word pairs), the MCVAE outperforms previous state-of-the-art models, such as Siamese-, CAE- and VAE-based RNNs." @default.
- W3110761489 created "2020-12-21" @default.
- W3110761489 creator A5015602781 @default.
- W3110761489 creator A5040305929 @default.
- W3110761489 creator A5075735963 @default.
- W3110761489 date "2020-12-03" @default.
- W3110761489 modified "2023-09-27" @default.
- W3110761489 title "A Correspondence Variational Autoencoder for Unsupervised Acoustic Word Embeddings." @default.
- W3110761489 cites W148837159 @default.
- W3110761489 cites W1959608418 @default.
- W3110761489 cites W2057007397 @default.
- W3110761489 cites W2059652594 @default.
- W3110761489 cites W2091746061 @default.
- W3110761489 cites W2114347655 @default.
- W3110761489 cites W2145410271 @default.
- W3110761489 cites W2157331557 @default.
- W3110761489 cites W2171590421 @default.
- W3110761489 cites W2190506272 @default.
- W3110761489 cites W2291770225 @default.
- W3110761489 cites W2295297373 @default.
- W3110761489 cites W2296681920 @default.
- W3110761489 cites W2407151108 @default.
- W3110761489 cites W2566587499 @default.
- W3110761489 cites W2753738274 @default.
- W3110761489 cites W2900308871 @default.
- W3110761489 cites W2932675979 @default.
- W3110761489 cites W2962736743 @default.
- W3110761489 cites W2962980711 @default.
- W3110761489 cites W2963223306 @default.
- W3110761489 cites W2963403868 @default.
- W3110761489 cites W2963425185 @default.
- W3110761489 cites W2963522047 @default.
- W3110761489 cites W2963720603 @default.
- W3110761489 cites W2963879199 @default.
- W3110761489 cites W2963879374 @default.
- W3110761489 cites W2964000524 @default.
- W3110761489 cites W2964121744 @default.
- W3110761489 cites W2964169922 @default.
- W3110761489 cites W2973096781 @default.
- W3110761489 cites W2990155616 @default.
- W3110761489 cites W3008444411 @default.
- W3110761489 cites W3015325583 @default.
- W3110761489 cites W3033010920 @default.
- W3110761489 cites W30845872 @default.
- W3110761489 cites W3131709275 @default.
- W3110761489 cites W3150750326 @default.
- W3110761489 cites W3162468843 @default.
- W3110761489 hasPublicationYear "2020" @default.
- W3110761489 type Work @default.
- W3110761489 sameAs 3110761489 @default.
- W3110761489 citedByCount "2" @default.
- W3110761489 countsByYear W31107614892020 @default.
- W3110761489 countsByYear W31107614892021 @default.
- W3110761489 crossrefType "posted-content" @default.
- W3110761489 hasAuthorship W3110761489A5015602781 @default.
- W3110761489 hasAuthorship W3110761489A5040305929 @default.
- W3110761489 hasAuthorship W3110761489A5075735963 @default.
- W3110761489 hasConcept C101738243 @default.
- W3110761489 hasConcept C146849305 @default.
- W3110761489 hasConcept C147168706 @default.
- W3110761489 hasConcept C153180895 @default.
- W3110761489 hasConcept C154945302 @default.
- W3110761489 hasConcept C17744445 @default.
- W3110761489 hasConcept C199539241 @default.
- W3110761489 hasConcept C204321447 @default.
- W3110761489 hasConcept C2524010 @default.
- W3110761489 hasConcept C2776359362 @default.
- W3110761489 hasConcept C2776436953 @default.
- W3110761489 hasConcept C33923547 @default.
- W3110761489 hasConcept C41008148 @default.
- W3110761489 hasConcept C50644808 @default.
- W3110761489 hasConcept C75165309 @default.
- W3110761489 hasConcept C8038995 @default.
- W3110761489 hasConcept C90805587 @default.
- W3110761489 hasConcept C94625758 @default.
- W3110761489 hasConceptScore W3110761489C101738243 @default.
- W3110761489 hasConceptScore W3110761489C146849305 @default.
- W3110761489 hasConceptScore W3110761489C147168706 @default.
- W3110761489 hasConceptScore W3110761489C153180895 @default.
- W3110761489 hasConceptScore W3110761489C154945302 @default.
- W3110761489 hasConceptScore W3110761489C17744445 @default.
- W3110761489 hasConceptScore W3110761489C199539241 @default.
- W3110761489 hasConceptScore W3110761489C204321447 @default.
- W3110761489 hasConceptScore W3110761489C2524010 @default.
- W3110761489 hasConceptScore W3110761489C2776359362 @default.
- W3110761489 hasConceptScore W3110761489C2776436953 @default.
- W3110761489 hasConceptScore W3110761489C33923547 @default.
- W3110761489 hasConceptScore W3110761489C41008148 @default.
- W3110761489 hasConceptScore W3110761489C50644808 @default.
- W3110761489 hasConceptScore W3110761489C75165309 @default.
- W3110761489 hasConceptScore W3110761489C8038995 @default.
- W3110761489 hasConceptScore W3110761489C90805587 @default.
- W3110761489 hasConceptScore W3110761489C94625758 @default.
- W3110761489 hasLocation W31107614891 @default.
- W3110761489 hasOpenAccess W3110761489 @default.
- W3110761489 hasPrimaryLocation W31107614891 @default.
- W3110761489 hasRelatedWork W2108665656 @default.
- W3110761489 hasRelatedWork W2116502739 @default.
- W3110761489 hasRelatedWork W2518754566 @default.
- W3110761489 hasRelatedWork W2795246076 @default.