FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2520176975> ?p ?o ?g. }

• W2520176975 endingPage "459" @default.
• W2520176975 startingPage "448" @default.
• W2520176975 abstract "In this paper, we present a unified approach to transfer learning of deep neural networks (DNNs) to address performance degradation issues caused by a potential acoustic mismatch between the training and testing conditions due to inter-speaker variability in state-of-the-art connectionist (a.k.a., hybrid) automatic speech recognition (ASR) systems. Different schemes to transfer knowledge of deep neural networks related to speaker adaptation can be developed with ease under such a unifying concept as demonstrated in the three frameworks investigated in this study. In the first solution, knowledge is transferred between homogeneous domains, namely the source and the target domains. Moreover the transfer takes place in a sequential manner from the target to the source speaker to boost the ASR accuracy on spoken utterances from a surprise target speaker. In the second solution, a multi-task approach is adopted to adjust the connectionist parameters to improve the ASR system performance on the target speaker. Knowledge is transferred simultaneously among heterogeneous tasks, and that is achieved by adding one or more smaller auxiliary output layers to the original DNN structure. In the third solution, DNN output classes are organised into a hierarchical structure in order to adjust the connectionist parameters and close the gap between training and testing conditions by transferring prior knowledge from the root node to the leaves in a structural maximum a posteriori fashion. Through a series of experiments on the Wall Street Journal (WSJ) speech recognition task, we show that the proposed solutions result in consistent and statistically significant word error rate reductions. Most importantly, we show that transfer learning is an enabling technology for speaker adaptation, since it outperforms both the transformation-based adaptation algorithms usually adapted in the speech community, and the multi-condition training (MCT) schemes, which is a data combination methods often adopted to cover more acoustic variabilities in speech when data from the source and target domains are both available at the training time. Finally, experimental evidence demonstrates that all proposed solutions are robust to negative transfer even when only a single sentence from the target speaker is available." @default.
• W2520176975 created "2016-09-23" @default.
• W2520176975 creator A5011008665 @default.
• W2520176975 creator A5066868860 @default.
• W2520176975 creator A5079659476 @default.
• W2520176975 date "2016-12-01" @default.
• W2520176975 modified "2023-10-16" @default.
• W2520176975 title "A unified approach to transfer learning of deep neural networks with applications to speaker adaptation in automatic speech recognition" @default.
• W2520176975 cites W1513862252 @default.
• W2520176975 cites W173856587 @default.
• W2520176975 cites W1964863507 @default.
• W2520176975 cites W1967558850 @default.
• W2520176975 cites W1980850109 @default.
• W2520176975 cites W1989549063 @default.
• W2520176975 cites W1992475611 @default.
• W2520176975 cites W2002342963 @default.
• W2520176975 cites W2024490156 @default.
• W2520176975 cites W2026369565 @default.
• W2520176975 cites W2033436836 @default.
• W2520176975 cites W2058641082 @default.
• W2520176975 cites W2060277733 @default.
• W2520176975 cites W2076794394 @default.
• W2520176975 cites W2079623482 @default.
• W2520176975 cites W2080005694 @default.
• W2520176975 cites W2094035326 @default.
• W2520176975 cites W2097685957 @default.
• W2520176975 cites W2100966557 @default.
• W2520176975 cites W2100969003 @default.
• W2520176975 cites W2106051978 @default.
• W2520176975 cites W2112021726 @default.
• W2520176975 cites W2118497033 @default.
• W2520176975 cites W2119203697 @default.
• W2520176975 cites W2121178390 @default.
• W2520176975 cites W2121981798 @default.
• W2520176975 cites W2131342762 @default.
• W2520176975 cites W2133924365 @default.
• W2520176975 cites W2136922672 @default.
• W2520176975 cites W2147768505 @default.
• W2520176975 cites W2147794814 @default.
• W2520176975 cites W2149175990 @default.
• W2520176975 cites W2154226865 @default.
• W2520176975 cites W2158069733 @default.
• W2520176975 cites W2160373860 @default.
• W2520176975 cites W2160815625 @default.
• W2520176975 cites W2161742217 @default.
• W2520176975 cites W2165698076 @default.
• W2520176975 cites W2294543795 @default.
• W2520176975 cites W2394932179 @default.
• W2520176975 cites W2397226255 @default.
• W2520176975 cites W2402040300 @default.
• W2520176975 cites W2407793339 @default.
• W2520176975 cites W2964182776 @default.
• W2520176975 cites W3162418253 @default.
• W2520176975 cites W3209042722 @default.
• W2520176975 cites W82936479 @default.
• W2520176975 doi "https://doi.org/10.1016/j.neucom.2016.09.018" @default.
• W2520176975 hasPublicationYear "2016" @default.
• W2520176975 type Work @default.
• W2520176975 sameAs 2520176975 @default.
• W2520176975 citedByCount "48" @default.
• W2520176975 countsByYear W25201769752017 @default.
• W2520176975 countsByYear W25201769752018 @default.
• W2520176975 countsByYear W25201769752019 @default.
• W2520176975 countsByYear W25201769752020 @default.
• W2520176975 countsByYear W25201769752021 @default.
• W2520176975 countsByYear W25201769752022 @default.
• W2520176975 countsByYear W25201769752023 @default.
• W2520176975 crossrefType "journal-article" @default.
• W2520176975 hasAuthorship W2520176975A5011008665 @default.
• W2520176975 hasAuthorship W2520176975A5066868860 @default.
• W2520176975 hasAuthorship W2520176975A5079659476 @default.
• W2520176975 hasConcept C120665830 @default.
• W2520176975 hasConcept C121332964 @default.
• W2520176975 hasConcept C133892786 @default.
• W2520176975 hasConcept C139807058 @default.
• W2520176975 hasConcept C150899416 @default.
• W2520176975 hasConcept C153180895 @default.
• W2520176975 hasConcept C154945302 @default.
• W2520176975 hasConcept C162324750 @default.
• W2520176975 hasConcept C187736073 @default.
• W2520176975 hasConcept C2780451532 @default.
• W2520176975 hasConcept C28490314 @default.
• W2520176975 hasConcept C2984842247 @default.
• W2520176975 hasConcept C41008148 @default.
• W2520176975 hasConcept C50644808 @default.
• W2520176975 hasConcept C8521452 @default.
• W2520176975 hasConceptScore W2520176975C120665830 @default.
• W2520176975 hasConceptScore W2520176975C121332964 @default.
• W2520176975 hasConceptScore W2520176975C133892786 @default.
• W2520176975 hasConceptScore W2520176975C139807058 @default.
• W2520176975 hasConceptScore W2520176975C150899416 @default.
• W2520176975 hasConceptScore W2520176975C153180895 @default.
• W2520176975 hasConceptScore W2520176975C154945302 @default.
• W2520176975 hasConceptScore W2520176975C162324750 @default.
• W2520176975 hasConceptScore W2520176975C187736073 @default.
• W2520176975 hasConceptScore W2520176975C2780451532 @default.
• W2520176975 hasConceptScore W2520176975C28490314 @default.
• W2520176975 hasConceptScore W2520176975C2984842247 @default.

• next