FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W4384822432> ?p ?o ?g. }

• W4384822432 endingPage "713" @default.
• W4384822432 startingPage "693" @default.
• W4384822432 abstract "Deep representation learning has gained significant momentum in advancing text-dependent speaker verification (TD-SV) systems. When designing deep neural networks (DNN) for extracting bottleneck (BN) features, the key considerations include training targets, activation functions, and loss functions. In this paper, we systematically study the impact of these choices on the performance of TD-SV. For training targets, we consider speaker identity, time-contrastive learning (TCL), and auto-regressive prediction coding, with the first being supervised and the last two being self-supervised. Furthermore, we study a range of loss functions when speaker identity is used as the training target. With regard to activation functions, we study the widely used sigmoid function, rectified linear unit (ReLU), and Gaussian error linear unit (GELU). We experimentally show that GELU is able to reduce the error rates of TD-SV significantly compared to sigmoid, irrespective of the training target. Among the three training targets, TCL performs the best. Among the various loss functions, cross-entropy, joint-softmax, and focal loss functions outperform the others. Finally, the score-level fusion of different systems is also able to reduce the error rates. To evaluate the representation learning methods, experiments are conducted on the RedDots 2016 challenge database consisting of short utterances for TD-SV systems based on classic Gaussian mixture model-universal background model (GMM-UBM) and i-vector methods." @default.
• W4384822432 created "2023-07-21" @default.
• W4384822432 creator A5002627413 @default.
• W4384822432 creator A5090108098 @default.
• W4384822432 date "2023-07-17" @default.
• W4384822432 modified "2023-10-14" @default.
• W4384822432 title "On Training Targets and Activation Functions for Deep Representation Learning in Text-Dependent Speaker Verification" @default.
• W4384822432 cites W1006777433 @default.
• W4384822432 cites W1528954144 @default.
• W4384822432 cites W1996512145 @default.
• W4384822432 cites W2035424729 @default.
• W4384822432 cites W2041823554 @default.
• W4384822432 cites W2046056978 @default.
• W4384822432 cites W2062227835 @default.
• W4384822432 cites W2090861223 @default.
• W4384822432 cites W2137075158 @default.
• W4384822432 cites W2148091317 @default.
• W4384822432 cites W2148154194 @default.
• W4384822432 cites W2150769028 @default.
• W4384822432 cites W2292259253 @default.
• W4384822432 cites W2326699523 @default.
• W4384822432 cites W2397474108 @default.
• W4384822432 cites W2491474862 @default.
• W4384822432 cites W2502631026 @default.
• W4384822432 cites W2508175215 @default.
• W4384822432 cites W2520774990 @default.
• W4384822432 cites W2550599163 @default.
• W4384822432 cites W2564171085 @default.
• W4384822432 cites W2888911591 @default.
• W4384822432 cites W2890964092 @default.
• W4384822432 cites W2900028034 @default.
• W4384822432 cites W2938358845 @default.
• W4384822432 cites W2949811029 @default.
• W4384822432 cites W2954930777 @default.
• W4384822432 cites W2963351448 @default.
• W4384822432 cites W2963466847 @default.
• W4384822432 cites W2963571336 @default.
• W4384822432 cites W2969985801 @default.
• W4384822432 cites W2973022971 @default.
• W4384822432 cites W2973066235 @default.
• W4384822432 cites W3016011332 @default.
• W4384822432 cites W3023641576 @default.
• W4384822432 cites W3094861110 @default.
• W4384822432 cites W3097796176 @default.
• W4384822432 cites W3099206234 @default.
• W4384822432 cites W3112869009 @default.
• W4384822432 cites W3126776772 @default.
• W4384822432 cites W3160804292 @default.
• W4384822432 cites W3161294170 @default.
• W4384822432 cites W3209059054 @default.
• W4384822432 cites W4224924217 @default.
• W4384822432 doi "https://doi.org/10.3390/acoustics5030042" @default.
• W4384822432 hasPublicationYear "2023" @default.
• W4384822432 type Work @default.
• W4384822432 citedByCount "1" @default.
• W4384822432 countsByYear W43848224322023 @default.
• W4384822432 crossrefType "journal-article" @default.
• W4384822432 hasAuthorship W4384822432A5002627413 @default.
• W4384822432 hasAuthorship W4384822432A5090108098 @default.
• W4384822432 hasBestOaLocation W43848224321 @default.
• W4384822432 hasConcept C108583219 @default.
• W4384822432 hasConcept C133892786 @default.
• W4384822432 hasConcept C153180895 @default.
• W4384822432 hasConcept C154945302 @default.
• W4384822432 hasConcept C167981619 @default.
• W4384822432 hasConcept C17744445 @default.
• W4384822432 hasConcept C188441871 @default.
• W4384822432 hasConcept C199539241 @default.
• W4384822432 hasConcept C2776359362 @default.
• W4384822432 hasConcept C28490314 @default.
• W4384822432 hasConcept C38365724 @default.
• W4384822432 hasConcept C40969351 @default.
• W4384822432 hasConcept C41008148 @default.
• W4384822432 hasConcept C50644808 @default.
• W4384822432 hasConcept C81388566 @default.
• W4384822432 hasConcept C94625758 @default.
• W4384822432 hasConceptScore W4384822432C108583219 @default.
• W4384822432 hasConceptScore W4384822432C133892786 @default.
• W4384822432 hasConceptScore W4384822432C153180895 @default.
• W4384822432 hasConceptScore W4384822432C154945302 @default.
• W4384822432 hasConceptScore W4384822432C167981619 @default.
• W4384822432 hasConceptScore W4384822432C17744445 @default.
• W4384822432 hasConceptScore W4384822432C188441871 @default.
• W4384822432 hasConceptScore W4384822432C199539241 @default.
• W4384822432 hasConceptScore W4384822432C2776359362 @default.
• W4384822432 hasConceptScore W4384822432C28490314 @default.
• W4384822432 hasConceptScore W4384822432C38365724 @default.
• W4384822432 hasConceptScore W4384822432C40969351 @default.
• W4384822432 hasConceptScore W4384822432C41008148 @default.
• W4384822432 hasConceptScore W4384822432C50644808 @default.
• W4384822432 hasConceptScore W4384822432C81388566 @default.
• W4384822432 hasConceptScore W4384822432C94625758 @default.
• W4384822432 hasIssue "3" @default.
• W4384822432 hasLocation W43848224321 @default.
• W4384822432 hasLocation W43848224322 @default.
• W4384822432 hasOpenAccess W4384822432 @default.
• W4384822432 hasPrimaryLocation W43848224321 @default.
• W4384822432 hasRelatedWork W2767072113 @default.

• next