generative model 0.0023825919999999998
true model 0.002327069
model states 0.00230975
mapping model 0.002289858
dependency model 0.002246133
model biases 0.00222885
model approximation 0.002228422
model prediction 0.0022252319999999997
model family 0.00219136
complex model 0.002190724
training data 0.00218154
model families 0.002173105
model change 0.00217189
model fami 0.002157591
ibm model 0.002157591
model predictions 0.002157591
model 0.00193517
hmm data 0.0017859500000000001
labeled data 0.0016088410000000002
true data 0.001561819
generative training 0.001459042
new training 0.001454475
much data 0.0014371520000000001
unsupervised models 0.0014172030000000001
real data 0.0014170390000000001
wsj data 0.001416354
artificial data 0.001404422
data increases 0.001396914
discriminative training 0.001395119
infinite data 0.0013933770000000001
training example 0.001364305
mixture models 0.001343015
training examples 0.001337037
same distribution 0.0012898760000000001
syntactic models 0.001280689
training procedure 0.001264714
same hmm 0.0012122880000000002
various models 0.001208347
different parameter 0.001184535
recursive models 0.0011748750000000001
dmv models 0.001162576
word distribution 0.001157839
different way 0.0011224389999999998
estimation error 0.001097946
local maximum 0.001042463
overall distribution 0.001039344
standard approach 0.001016291
training 0.00101162
true hmm 0.001007929
language processing 0.001001739
local search 9.91156E-4
pos tags 9.876659999999999E-4
language acquisition 9.84455E-4
other emission 9.8364E-4
objective function 9.681760000000001E-4
many methods 9.652630000000001E-4
same value 9.59313E-4
unsupervised grammar 9.59233E-4
local optima 9.539780000000001E-4
stopword distribution 9.53076E-4
optimization error 9.51328E-4
initial approach 9.34472E-4
likelihood objective 9.2391E-4
models 9.23692E-4
ity distribution 9.22411E-4
density estimation 9.19696E-4
conditional likelihood 9.181829999999999E-4
hmm migrations 9.17986E-4
first iteration 9.04041E-4
objective way 8.8903E-4
empirical likelihood 8.88393E-4
joint likelihood 8.85232E-4
linguistic structure 8.84571E-4
corresponding sequence 8.81518E-4
induction accuracy 8.76824E-4
true parameters 8.686080000000001E-4
unsupervised learning 8.66715E-4
top hmm 8.643520000000001E-4
parameter setting 8.587900000000001E-4
parameter settings 8.555430000000001E-4
unsupervised case 8.52217E-4
same phenomenon 8.521690000000001E-4
optimal parameter 8.46891E-4
same part 8.46555E-4
unsupervised induction 8.39864E-4
hmm migra 8.393140000000001E-4
estimation scenarios 8.348089999999999E-4
new objective 8.347739999999999E-4
treebank tags 8.34266E-4
unlabeled examples 8.34141E-4
tify estimation 8.32798E-4
prediction accuracy 8.20533E-4
same raising 8.19733E-4
same probabil 8.19733E-4
first observation 8.14367E-4
stop probability 8.130520000000001E-4
grammar induction 8.12075E-4
ing examples 8.10484E-4
many jjs 8.09356E-4
many authors 8.09356E-4
