hmm state 0.00374514
state hmm 0.00374514
hmm states 0.0027919900000000003
different hmm 0.00272225
state split 0.0025941880000000003
single state 0.002564325
state hmms 0.002457987
successive state 0.002437672
original state 0.002427301
state replacement 0.00242114
state occupancy 0.0023957180000000002
state splitting 0.002378453
concurrent state 0.00237469
aggressive state 0.00237469
modified state 0.00237469
hmm label 0.002131783
state 0.00211872
new hmm 0.0020844419999999997
markov model 0.00204943
hmm size 0.00201693
model parameters 0.001998115
current hmm 0.001953308
hmm sizes 0.0019362749999999999
separate hmm 0.001933236
hmm cor 0.001928953
the hmm 0.001916266
hmm setting 0.001903941
hmm labeler 0.001882234
training speech 0.001864677
other states 0.001847872
bigram model 0.001815841
training set 0.0016949909999999999
speech data 0.001630884
states sequences 0.0015788960000000002
model 0.00151918
states sss 0.0014788840000000002
original states 0.0014741510000000002
distinct states 0.001429057
ferent states 0.001428412
phone sequence 0.0014256429999999999
phone accuracy 0.00139394
different allophones 0.001376226
different ways 0.0013627449999999998
different amounts 0.001354919
ing speech 0.001349856
speech segment 0.001310498
speech recognition 0.001262549
acoustic units 0.001215735
phone segment 0.0012104149999999998
label sequence 0.001209862
automatic speech 0.001172011
unsupervised learning 0.0011711410000000001
speech signal 0.001167126
test speech 0.001165925
states 0.00116557
phone recognition 0.001162466
statistical models 0.001143238
phone sequences 0.0011344699999999998
japanese speech 0.001119082
recognition accuracy 0.001114118
standard algorithm 0.001094978
speech sounds 0.001083672
unlabeled speech 0.001080433
experimental results 0.001059731
allophonic models 0.001058865
training 0.00104345
pronunciation lexicons 0.001033935
bigram models 0.001025176
likely phone 0.001009011
hidden markov 9.97952E-4
test set 9.96239E-4
ditional models 9.96047E-4
phonemic models 9.913629999999999E-4
viterbi algorithm 9.817439999999999E-4
phone loop 9.7971E-4
phonemic pronunciation 9.51542E-4
phoneme set 9.455869999999999E-4
performance accuracies 9.44019E-4
sss algorithm 9.387169999999999E-4
automatic learning 9.260469999999999E-4
first column 9.25578E-4
same iteration 9.22951E-4
label sequences 9.18689E-4
first experiment 9.08827E-4
ent learning 8.82199E-4
learning experiments 8.50171E-4
early work 8.39102E-4
gaussian density 8.38941E-4
gaussian densities 8.358459999999999E-4
learning setups 8.3317E-4
standard viterbi 8.25916E-4
speech 8.21227E-4
phonetic transcription 8.04535E-4
sample mean 7.85339E-4
evaluation methodology 7.76168E-4
units the 7.71587E-4
symbolic units 7.66194E-4
simple bigram 7.55302E-4
linguistic knowledge 7.5086E-4
temporal split 7.47287E-4
