FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2108672476> ?p ?o ?g. }

• W2108672476 endingPage "311" @default.
• W2108672476 startingPage "299" @default.
• W2108672476 abstract "Hidden markov models (HMM's) are one of the most powerful speech recognition tools available today. Even so, the inadequacies of HMM's as a correct modeling framework for speech are well known. In this context, it is argued in this paper that the maximum mutual information estimation (MMIE) formulation for training is more appropriate than maximum likelihood estimation (MLE) for reducing the error rate. Corrective MMIE training is introduced. It is a very efficient new training algorithm which uses a modified version of a discrete reestimation formula recently proposed by Gopalakrishnan et al.( see IEEE Trans. Inform. Theory, Jan. 1991). Reestimation formulas are proposed for the case of diagonal Gaussian densities and their convergence properties are experimentally demonstrated. A description of how these formulas are integrated into our training algorithm is given. Using the MMIE framework for training, it is shown how weighting the contribution of different parameter sets in the computation of output probabilities introduces substantial recognition improvements. Using the TIDIGITS connected digit corpus, a large number of experiments are performed with the ideas, techniques, and algorithms presented in this paper. These experiments show that MMIE systematically provides substantial error rate reductions with respect to MLE alone and that, thanks to the new training techniques, these results can be obtained at an acceptable computational cost. The best results obtained in the experiments were 0.29% word error rate and 0.89% string error rate on the adult portion of the corpus.< <ETX xmlns:mml=http://www.w3.org/1998/Math/MathML xmlns:xlink=http://www.w3.org/1999/xlink>></ETX>" @default.
• W2108672476 created "2016-06-24" @default.
• W2108672476 creator A5065257419 @default.
• W2108672476 creator A5065576817 @default.
• W2108672476 creator A5081408760 @default.
• W2108672476 date "1994-04-01" @default.
• W2108672476 modified "2023-10-14" @default.
• W2108672476 title "High-performance connected digit recognition using maximum mutual information estimation" @default.
• W2108672476 cites W1524975733 @default.
• W2108672476 cites W1538748823 @default.
• W2108672476 cites W1562142077 @default.
• W2108672476 cites W1865845602 @default.
• W2108672476 cites W1877570817 @default.
• W2108672476 cites W1884432169 @default.
• W2108672476 cites W1975953721 @default.
• W2108672476 cites W2039151035 @default.
• W2108672476 cites W2063311813 @default.
• W2108672476 cites W2077574412 @default.
• W2108672476 cites W2098601596 @default.
• W2108672476 cites W2101858445 @default.
• W2108672476 cites W2102804087 @default.
• W2108672476 cites W2115002762 @default.
• W2108672476 cites W2152131029 @default.
• W2108672476 cites W2155900437 @default.
• W2108672476 cites W2170585560 @default.
• W2108672476 cites W2171244697 @default.
• W2108672476 cites W2945307304 @default.
• W2108672476 cites W4239340178 @default.
• W2108672476 cites W4361866991 @default.
• W2108672476 doi "https://doi.org/10.1109/89.279279" @default.
• W2108672476 hasPublicationYear "1994" @default.
• W2108672476 type Work @default.
• W2108672476 sameAs 2108672476 @default.
• W2108672476 citedByCount "102" @default.
• W2108672476 countsByYear W21086724762012 @default.
• W2108672476 countsByYear W21086724762013 @default.
• W2108672476 countsByYear W21086724762014 @default.
• W2108672476 countsByYear W21086724762015 @default.
• W2108672476 countsByYear W21086724762016 @default.
• W2108672476 countsByYear W21086724762020 @default.
• W2108672476 countsByYear W21086724762021 @default.
• W2108672476 countsByYear W21086724762022 @default.
• W2108672476 crossrefType "journal-article" @default.
• W2108672476 hasAuthorship W2108672476A5065257419 @default.
• W2108672476 hasAuthorship W2108672476A5065576817 @default.
• W2108672476 hasAuthorship W2108672476A5081408760 @default.
• W2108672476 hasConcept C11413529 @default.
• W2108672476 hasConcept C121332964 @default.
• W2108672476 hasConcept C126838900 @default.
• W2108672476 hasConcept C151730666 @default.
• W2108672476 hasConcept C152139883 @default.
• W2108672476 hasConcept C153180895 @default.
• W2108672476 hasConcept C154945302 @default.
• W2108672476 hasConcept C157486923 @default.
• W2108672476 hasConcept C162324750 @default.
• W2108672476 hasConcept C163716315 @default.
• W2108672476 hasConcept C183115368 @default.
• W2108672476 hasConcept C23224414 @default.
• W2108672476 hasConcept C2524010 @default.
• W2108672476 hasConcept C2777303404 @default.
• W2108672476 hasConcept C2779343474 @default.
• W2108672476 hasConcept C28490314 @default.
• W2108672476 hasConcept C33923547 @default.
• W2108672476 hasConcept C37914503 @default.
• W2108672476 hasConcept C40969351 @default.
• W2108672476 hasConcept C41008148 @default.
• W2108672476 hasConcept C45374587 @default.
• W2108672476 hasConcept C50522688 @default.
• W2108672476 hasConcept C62520636 @default.
• W2108672476 hasConcept C71924100 @default.
• W2108672476 hasConcept C86803240 @default.
• W2108672476 hasConcept C90805587 @default.
• W2108672476 hasConceptScore W2108672476C11413529 @default.
• W2108672476 hasConceptScore W2108672476C121332964 @default.
• W2108672476 hasConceptScore W2108672476C126838900 @default.
• W2108672476 hasConceptScore W2108672476C151730666 @default.
• W2108672476 hasConceptScore W2108672476C152139883 @default.
• W2108672476 hasConceptScore W2108672476C153180895 @default.
• W2108672476 hasConceptScore W2108672476C154945302 @default.
• W2108672476 hasConceptScore W2108672476C157486923 @default.
• W2108672476 hasConceptScore W2108672476C162324750 @default.
• W2108672476 hasConceptScore W2108672476C163716315 @default.
• W2108672476 hasConceptScore W2108672476C183115368 @default.
• W2108672476 hasConceptScore W2108672476C23224414 @default.
• W2108672476 hasConceptScore W2108672476C2524010 @default.
• W2108672476 hasConceptScore W2108672476C2777303404 @default.
• W2108672476 hasConceptScore W2108672476C2779343474 @default.
• W2108672476 hasConceptScore W2108672476C28490314 @default.
• W2108672476 hasConceptScore W2108672476C33923547 @default.
• W2108672476 hasConceptScore W2108672476C37914503 @default.
• W2108672476 hasConceptScore W2108672476C40969351 @default.
• W2108672476 hasConceptScore W2108672476C41008148 @default.
• W2108672476 hasConceptScore W2108672476C45374587 @default.
• W2108672476 hasConceptScore W2108672476C50522688 @default.
• W2108672476 hasConceptScore W2108672476C62520636 @default.
• W2108672476 hasConceptScore W2108672476C71924100 @default.
• W2108672476 hasConceptScore W2108672476C86803240 @default.
• W2108672476 hasConceptScore W2108672476C90805587 @default.

• next