TY  - GEN
AB  - In  current  large  vocabulary  continuous  speech  recognition  systems,  multivariate  Gaussian  mixture  distributions  and  context-dependent  phones,  typically  triphones,  are  used to achieve high  accuracy  acoustic models.  It is crucial to address the problem of  how  to estimate an extremely large number of  model parameters from a limited  amount of  training data.  The traditional approach uses  phonetic  decision tree based  context  clustering for  reducing  free  parameters. However,  this  approach  has several problems  that  might cause  system  performance  degradation. All  of these  problems  are  due  to the  fact  that  the traditional  approach  does  not  efficiently  use  the limited  training data and  therefore  fails to obtain effective acoustic models.
AD  - Oregon Health and Science University
AU  - Liu, Chaojun
DA  - 2002-06-01
DO  - 10.6083/M43F4MX9
DO  - DOI
ED  - Yan, Yonghong
ED  - Advisor
ID  - 2479
KW  - Phonetics
KW  - Speech Recognition Software
L1  - https://digitalcollections.ohsu.edu/record/2479/files/3220_etd.pdf
L2  - https://digitalcollections.ohsu.edu/record/2479/files/3220_etd.pdf
L4  - https://digitalcollections.ohsu.edu/record/2479/files/3220_etd.pdf
LK  - https://digitalcollections.ohsu.edu/record/2479/files/3220_etd.pdf
N2  - In  current  large  vocabulary  continuous  speech  recognition  systems,  multivariate  Gaussian  mixture  distributions  and  context-dependent  phones,  typically  triphones,  are  used to achieve high  accuracy  acoustic models.  It is crucial to address the problem of  how  to estimate an extremely large number of  model parameters from a limited  amount of  training data.  The traditional approach uses  phonetic  decision tree based  context  clustering for  reducing  free  parameters. However,  this  approach  has several problems  that  might cause  system  performance  degradation. All  of these  problems  are  due  to the  fact  that  the traditional  approach  does  not  efficiently  use  the limited  training data and  therefore  fails to obtain effective acoustic models.
PB  - Oregon Health and Sciences University
PY  - 2002-06-01
T1  - Toward more effective acoustic model clustering by more efficient use of data in speech recognition
TI  - Toward more effective acoustic model clustering by more efficient use of data in speech recognition
UR  - https://digitalcollections.ohsu.edu/record/2479/files/3220_etd.pdf
Y1  - 2002-06-01
ER  -