model features 0.0035399
word pitch 0.003026022
sequence model 0.002918556
first word 0.0028113440000000003
previous word 0.002793901
function word 0.002790331
following word 0.0027399480000000003
word frequency 0.0027358630000000003
ing word 0.002712135
current word 0.002693071
word level 0.0026602600000000002
target word 0.002650855
last word 0.002635332
preceding word 0.002610074
word fre 0.0026098420000000002
model pitch 0.002594372
model label 0.002529748
model accuracy 0.0024929929999999998
markov model 0.002464578
crf model 0.002451596
hmm model 0.002449485
prediction model 0.002295347
same features 0.002291086
intonational model 0.002252766
other words 0.00223629
quence model 0.002217803
discriminative model 0.0022125499999999998
complete model 0.00217076
previous words 0.002157211
function words 0.002153641
content words 0.002010544
surrounding words 0.00199212
accented words 0.0019883
textual features 0.001971155
acoustic features 0.001963225
model 0.00192433
phonological features 0.001909235
tic features 0.001907871
interacting features 0.0018719309999999999
tactic features 0.001868697
suprasegmental features 0.001861268
syntactic information 0.00175869
words 0.00171929
speech recognition 0.0016234220000000002
features 0.00161557
label sequence 0.001599644
feature representation 0.0015690349999999999
feature extraction 0.0015423169999999999
log speech 0.001517548
feature repre 0.001509503
similar information 0.001500125
speech boundaries 0.001494471
acoustic data 0.001408815
textual information 0.001399805
acoustic information 0.0013918749999999999
conversational speech 0.001388625
speech rate 0.001380293
labeled data 0.0013694290000000001
pitch accent 0.001348405
speech synthesis 0.0013377550000000002
information content 0.0013354740000000001
contextual information 0.0013330899999999999
critical information 0.001331825
the data 0.001328232
spontaneous speech 0.001325636
mutual information 0.001316439
noisy data 0.001307315
speech relay 0.0013039990000000001
observation sequence 0.001302315
sequence labeling 0.001296384
tomatic speech 0.001294442
information sun 0.001290331
information effi 0.001290331
training label 0.0012891579999999999
speech synthe 0.0012870280000000002
markov models 0.001273534
sequence pair 0.0012704769999999999
test accuracy 0.0012656809999999998
feature 0.0012632
likely sequence 0.0012534809999999999
bel sequence 0.001240178
sequence perceptron 0.001240178
vation sequence 0.001240178
sequence adaboost 0.001240178
label accuracy 0.001174081
previous models 0.001171207
syntactic variables 0.001157365
conditional probability 0.0011545190000000001
syntactic category 0.0011347800000000002
modeling pitch 0.001128501
same variables 0.0011184110000000001
prediction models 0.001104303
different speakers 0.0010952409999999998
high probability 0.0010816509999999999
acoustic models 0.001080941
different sources 0.001072682
joint probability 0.0010641560000000001
accent prediction 0.00104938
information 0.00104422
speech 0.00104121
