model training 0.003167355
prosodic model 0.0030896
lexical model 0.002928855
language model 0.00290931
training data 0.0027457849999999997
prosody model 0.002714581
acoustic model 0.0025919940000000002
different data 0.00258168
model scores 0.002539786
model likelihood 0.0025147950000000002
observation model 0.002496574
prosodic features 0.00248796
expectation model 0.002484481
pronunciation model 0.0024825940000000003
guage model 0.002467023
tic model 0.002448496
proved model 0.002443226
model behavior 0.002443226
data corpus 0.002380233
test data 0.002334244
lexical features 0.0023272149999999997
data set 0.002301865
first data 0.002265714
news data 0.002262123
model 0.00219225
prosodic models 0.00214946
data sets 0.002099672
word error 0.002091865
unlabeled data 0.0020867389999999998
new data 0.002085879
second data 0.002085782
data mismatch 0.002049478
journal data 0.002046644
labeled data 0.002037431
development data 0.002035321
entire data 0.002034509
data partition 0.002024624
duce data 0.002023633
lexical models 0.0019887150000000003
language models 0.00196917
speech recognition 0.001924123
detailed features 0.001915764
final features 0.001903356
duration features 0.0018568489999999998
delta features 0.0018482149999999998
spectral features 0.001843836
suprasegmental features 0.001842111
word sequence 0.001808817
other training 0.0017759429999999999
prosody models 0.001774441
prosodic information 0.001767939
word boundary 0.001765455
news speech 0.001724793
binary feature 0.001710816
speech domain 0.0016677250000000001
acoustic models 0.0016518540000000001
term models 0.0016477780000000001
word sequences 0.0016460560000000001
word output 0.001617759
lexical information 0.001607194
candidate word 0.0016004160000000001
features 0.00159061
training corpus 0.001584658
sri speech 0.001562005
automatic speech 0.001557115
proper word 0.0015527790000000002
feature extraction 0.001551654
erroneous word 0.0015381230000000002
guage models 0.0015268830000000001
speech recognizer 0.001514592
spontaneous speech 0.001507418
training set 0.00150629
speech understanding 0.001505197
tection models 0.001504845
read speech 0.00150223
different recognition 0.001501773
speech signal 0.00150012
ter speech 0.0014992080000000001
recognition error 0.0014961380000000002
pos tag 0.0014896739999999999
speech recogni 0.0014887960000000001
speech researchers 0.001487546
art speech 0.001485122
ural speech 0.001485122
speech waveform 0.001485122
different asr 0.001478685
recognition performance 0.001448836
lexical pos 0.001439683
asr performance 0.0014257480000000001
correct words 0.00141216
error detection 0.001383218
boundary information 0.0013495439999999998
detection performance 0.001335916
supervised training 0.001333771
prosodic labels 0.001311646
training methods 0.001306518
prosodic knowledge 0.001302048
several pos 0.001301788
bold words 0.001294643
prosodic event 0.001293629
