FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2094147890> ?p ?o ?g. }

• W2094147890 abstract "This paper proposes a simple yet effective model-based neural network speaker adaptation technique that learns speaker-specific hidden unit contributions given adaptation data, without requiring any form of speaker-adaptive training, or labelled adaptation data. An additional amplitude parameter is defined for each hidden unit; the amplitude parameters are tied for each speaker, and are learned using unsupervised adaptation. We conducted experiments on the TED talks data, as used in the International Workshop on Spoken Language Translation (IWSLT) evaluations. Our results indicate that the approach can reduce word error rates on standard IWSLT test sets by about 8-15% relative compared to unadapted systems, with a further reduction of 4-6% relative when combined with feature-space maximum likelihood linear regression (fMLLR). The approach can be employed in most existing feed-forward neural network architectures, and we report results using various hidden unit activation functions: sigmoid, maxout, and rectifying linear units (ReLU)." @default.
• W2094147890 created "2016-06-24" @default.
• W2094147890 creator A5083001976 @default.
• W2094147890 creator A5084169592 @default.
• W2094147890 date "2014-12-01" @default.
• W2094147890 modified "2023-10-16" @default.
• W2094147890 title "Learning hidden unit contributions for unsupervised speaker adaptation of neural network acoustic models" @default.
• W2094147890 cites W1513862252 @default.
• W2094147890 cites W1537275613 @default.
• W2094147890 cites W1968706335 @default.
• W2094147890 cites W1985368711 @default.
• W2094147890 cites W1989549063 @default.
• W2094147890 cites W1993409002 @default.
• W2094147890 cites W2002342963 @default.
• W2094147890 cites W2010362084 @default.
• W2094147890 cites W2012897754 @default.
• W2094147890 cites W2015633636 @default.
• W2094147890 cites W2036351241 @default.
• W2094147890 cites W2056738732 @default.
• W2094147890 cites W2056825827 @default.
• W2094147890 cites W2060277733 @default.
• W2094147890 cites W2061305730 @default.
• W2094147890 cites W2062164080 @default.
• W2094147890 cites W2079623482 @default.
• W2094147890 cites W2080005694 @default.
• W2094147890 cites W2087006792 @default.
• W2094147890 cites W2105099419 @default.
• W2094147890 cites W2119203697 @default.
• W2094147890 cites W2146871184 @default.
• W2094147890 cites W2149194912 @default.
• W2094147890 cites W2150769028 @default.
• W2094147890 cites W2160306971 @default.
• W2094147890 cites W2160650576 @default.
• W2094147890 cites W2160815625 @default.
• W2094147890 cites W2164240571 @default.
• W2094147890 cites W2165712214 @default.
• W2094147890 cites W2169189000 @default.
• W2094147890 cites W2172130635 @default.
• W2094147890 cites W2484208911 @default.
• W2094147890 cites W811578723 @default.
• W2094147890 doi "https://doi.org/10.1109/slt.2014.7078569" @default.
• W2094147890 hasPublicationYear "2014" @default.
• W2094147890 type Work @default.
• W2094147890 sameAs 2094147890 @default.
• W2094147890 citedByCount "195" @default.
• W2094147890 countsByYear W20941478902014 @default.
• W2094147890 countsByYear W20941478902015 @default.
• W2094147890 countsByYear W20941478902016 @default.
• W2094147890 countsByYear W20941478902017 @default.
• W2094147890 countsByYear W20941478902018 @default.
• W2094147890 countsByYear W20941478902019 @default.
• W2094147890 countsByYear W20941478902020 @default.
• W2094147890 countsByYear W20941478902021 @default.
• W2094147890 countsByYear W20941478902022 @default.
• W2094147890 countsByYear W20941478902023 @default.
• W2094147890 crossrefType "proceedings-article" @default.
• W2094147890 hasAuthorship W2094147890A5083001976 @default.
• W2094147890 hasAuthorship W2094147890A5084169592 @default.
• W2094147890 hasBestOaLocation W20941478902 @default.
• W2094147890 hasConcept C119857082 @default.
• W2094147890 hasConcept C120665830 @default.
• W2094147890 hasConcept C121332964 @default.
• W2094147890 hasConcept C133892786 @default.
• W2094147890 hasConcept C138885662 @default.
• W2094147890 hasConcept C139807058 @default.
• W2094147890 hasConcept C153180895 @default.
• W2094147890 hasConcept C154945302 @default.
• W2094147890 hasConcept C16910744 @default.
• W2094147890 hasConcept C199360897 @default.
• W2094147890 hasConcept C23224414 @default.
• W2094147890 hasConcept C2776401178 @default.
• W2094147890 hasConcept C28490314 @default.
• W2094147890 hasConcept C41008148 @default.
• W2094147890 hasConcept C41895202 @default.
• W2094147890 hasConcept C50644808 @default.
• W2094147890 hasConcept C81388566 @default.
• W2094147890 hasConceptScore W2094147890C119857082 @default.
• W2094147890 hasConceptScore W2094147890C120665830 @default.
• W2094147890 hasConceptScore W2094147890C121332964 @default.
• W2094147890 hasConceptScore W2094147890C133892786 @default.
• W2094147890 hasConceptScore W2094147890C138885662 @default.
• W2094147890 hasConceptScore W2094147890C139807058 @default.
• W2094147890 hasConceptScore W2094147890C153180895 @default.
• W2094147890 hasConceptScore W2094147890C154945302 @default.
• W2094147890 hasConceptScore W2094147890C16910744 @default.
• W2094147890 hasConceptScore W2094147890C199360897 @default.
• W2094147890 hasConceptScore W2094147890C23224414 @default.
• W2094147890 hasConceptScore W2094147890C2776401178 @default.
• W2094147890 hasConceptScore W2094147890C28490314 @default.
• W2094147890 hasConceptScore W2094147890C41008148 @default.
• W2094147890 hasConceptScore W2094147890C41895202 @default.
• W2094147890 hasConceptScore W2094147890C50644808 @default.
• W2094147890 hasConceptScore W2094147890C81388566 @default.
• W2094147890 hasLocation W20941478901 @default.
• W2094147890 hasLocation W20941478902 @default.
• W2094147890 hasLocation W20941478903 @default.
• W2094147890 hasOpenAccess W2094147890 @default.
• W2094147890 hasPrimaryLocation W20941478901 @default.
• W2094147890 hasRelatedWork W1497807607 @default.
• W2094147890 hasRelatedWork W1509309911 @default.

• next