Matches in SemOpenAlex for { <https://semopenalex.org/work/W2079577056> ?p ?o ?g. }
Showing items 1 to 64 of
64
with 100 items per page.
- W2079577056 endingPage "17" @default.
- W2079577056 startingPage "5" @default.
- W2079577056 abstract "Abstract In this paper the design of accurate Semi-Continuous Density Hidden Markov Models (SC-HMMs) for acoustic modelling in large vocabulary continuous speech recognition is presented. Two methods are described to improve drastically the efficiency of the observation likelihood calculations for the SC-HMMs. First, reduced SC-HMMs are created, where each state does not share all the – gaussian – probability density functions ( pdfs ) but only those which are important for it. It is shown how the average number of gaussians per state can be reduced to 70 for a total set of 10 000 gaussians. Second, a novel scalar selection algorithm is presented reducing to 5% the number of gaussians which have to be calculated on the total set of 10 000, without any degradation in recognition performance. Furthermore, the concept of tied state context-dependent modelling with phonetic decision trees is adapted to SC-HMMs. In fact, a node splitting criterion appropriate for SC-HMMs is introduced: it is based on a distance measure between the mixtures of gaussian pdfs as involved in SC-HMM state modelling. This contrasts with other criteria from literature which are based on simplified pdfs to manage the algorithmic complexity. On the ARPA Resource Management task, a relative reduction in word error rate of 8% was achieved with the proposed criterion, comparing with two known criteria based on simplified pdfs ." @default.
- W2079577056 created "2016-06-24" @default.
- W2079577056 creator A5041039488 @default.
- W2079577056 creator A5046536366 @default.
- W2079577056 creator A5073887372 @default.
- W2079577056 date "1998-04-01" @default.
- W2079577056 modified "2023-09-24" @default.
- W2079577056 title "Fast and accurate acoustic modelling with semi-continuous HMMs" @default.
- W2079577056 cites W103520368 @default.
- W2079577056 cites W121156287 @default.
- W2079577056 cites W1530913040 @default.
- W2079577056 cites W1555410120 @default.
- W2079577056 cites W1592898276 @default.
- W2079577056 cites W1627936606 @default.
- W2079577056 cites W1819945869 @default.
- W2079577056 cites W1824013548 @default.
- W2079577056 cites W1900367739 @default.
- W2079577056 cites W2050597349 @default.
- W2079577056 cites W2075815989 @default.
- W2079577056 cites W2096466030 @default.
- W2079577056 cites W2098773974 @default.
- W2079577056 cites W2106420626 @default.
- W2079577056 cites W2121178390 @default.
- W2079577056 cites W2152494161 @default.
- W2079577056 cites W2170120894 @default.
- W2079577056 cites W2489815359 @default.
- W2079577056 cites W84269282 @default.
- W2079577056 doi "https://doi.org/10.1016/s0167-6393(98)00002-8" @default.
- W2079577056 hasPublicationYear "1998" @default.
- W2079577056 type Work @default.
- W2079577056 sameAs 2079577056 @default.
- W2079577056 citedByCount "37" @default.
- W2079577056 countsByYear W20795770562016 @default.
- W2079577056 crossrefType "journal-article" @default.
- W2079577056 hasAuthorship W2079577056A5041039488 @default.
- W2079577056 hasAuthorship W2079577056A5046536366 @default.
- W2079577056 hasAuthorship W2079577056A5073887372 @default.
- W2079577056 hasConcept C23224414 @default.
- W2079577056 hasConcept C28490314 @default.
- W2079577056 hasConcept C41008148 @default.
- W2079577056 hasConceptScore W2079577056C23224414 @default.
- W2079577056 hasConceptScore W2079577056C28490314 @default.
- W2079577056 hasConceptScore W2079577056C41008148 @default.
- W2079577056 hasIssue "1" @default.
- W2079577056 hasLocation W20795770561 @default.
- W2079577056 hasOpenAccess W2079577056 @default.
- W2079577056 hasPrimaryLocation W20795770561 @default.
- W2079577056 hasRelatedWork W1521297879 @default.
- W2079577056 hasRelatedWork W197415996 @default.
- W2079577056 hasRelatedWork W1979120075 @default.
- W2079577056 hasRelatedWork W2109705048 @default.
- W2079577056 hasRelatedWork W2131226854 @default.
- W2079577056 hasRelatedWork W2136763963 @default.
- W2079577056 hasRelatedWork W2374918184 @default.
- W2079577056 hasRelatedWork W2793122029 @default.
- W2079577056 hasRelatedWork W2940588515 @default.
- W2079577056 hasRelatedWork W3128571556 @default.
- W2079577056 hasVolume "24" @default.
- W2079577056 isParatext "false" @default.
- W2079577056 isRetracted "false" @default.
- W2079577056 magId "2079577056" @default.
- W2079577056 workType "article" @default.