unsupervised model 0.002383581
sequence model 0.002374142
syntactic model 0.00230083
generative model 0.002297409
model selection 0.002283628
grammatical model 0.002246447
oracle model 0.0022149879999999998
unigram model 0.002195515
tic model 0.002188864
model 0.001947
training data 0.001937086
different feature 0.001880397
other training 0.0018148359999999998
feature function 0.001765845
local features 0.001730381
same feature 0.001716819
feature set 0.001674729
training set 0.001666269
conditional training 0.001658955
language language 0.001608034
feature weight 0.001559354
vector feature 0.001556875
good models 0.001543449
length models 0.001538767
unsupervised models 0.0015295109999999999
sequence models 0.001520072
training example 0.001506336
probabilistic models 0.001505118
feature value 0.001458103
corresponding feature 0.001457649
supervised training 0.0014545489999999999
training examples 0.0014528429999999999
new features 0.0014449609999999998
generative models 0.001443339
feature weights 0.001438998
supervised models 0.001428689
linear models 0.001420835
feature types 0.001409161
training criteria 0.0013882719999999999
spelling features 0.001384508
margin training 0.001378069
jth feature 0.001376591
training products 0.0013748889999999998
training iterations 0.001371598
training exam 0.001357743
ith training 0.001356366
training cor 0.001349185
acoustic models 0.001332185
ical models 0.001324457
additional features 0.0013200389999999998
novel features 0.001315482
different set 0.001300626
trary features 0.001275945
new language 0.001202738
other sentence 0.001200687
natural language 0.001199915
local accuracy 0.001175765
development data 0.001143519
complete data 0.001139936
feature 0.00112725
training 0.00111879
unlabeled data 0.001115105
conditional estimation 0.001105272
dependency grammar 0.001104667
models 0.00109293
thing language 0.001089676
language syntax 0.001082065
local maximum 0.001077487
different criterion 0.00106939
bad data 0.001061172
other neighborhood 0.0010572799999999999
unannotated data 0.0010553140000000001
annotated data 0.001052641
parameter estimation 0.0010510950000000002
much data 0.001049253
features 0.00104624
language support 0.0010409949999999999
tag set 0.001035982
same length 0.001035406
dyna language 0.001033354
ural language 0.001033354
different sets 0.001028485
pos tag 0.0010233070000000002
first sentence 0.0010136820000000001
different ways 0.0010113420000000001
objective function 0.001010885
unsupervised estimation 0.0010016880000000001
different level 0.001001327
neighborhood function 9.99829E-4
different dilution 9.90146E-4
word types 9.86584E-4
different plots 9.86135E-4
different variances 9.85518E-4
good pos 9.85323E-4
unsupervised learning 9.85181E-4
different neighbor 9.83838E-4
probabilistic estimation 9.77295E-4
word order 9.75972E-4
other criteria 9.65528E-4
pos sequence 9.61946E-4
