training data 0.00336963
unlabeled data 0.003201674
test data 0.003129691
same data 0.002934863
labeled data 0.0029099269999999997
segmented data 0.002787328
development data 0.002722592
input data 0.002706535
experimental data 0.002702963
velopment data 0.002626275
word segmentation 0.00255687
joint model 0.002360136
cws model 0.002172062
crfs model 0.002151691
current model 0.002137294
chinese word 0.002129194
model induction 0.002103023
perceptrons model 0.002040124
everlasting model 0.002030318
possible word 0.002008645
word seg 0.00192662
segmentation models 0.001863179
word boundaries 0.001850832
explicit word 0.001824636
model 0.00176676
segmentation performance 0.0017595549999999999
feature vector 0.001737022
feature functions 0.001650582
different models 0.0015561210000000002
global feature 0.001534183
unlabeled sentences 0.001531025
feature templates 0.0015195249999999999
feature vec 0.001507735
arbitrary feature 0.001506143
feature engineering 0.0015028469999999999
supervised models 0.00148036
segmentation result 0.001446842
label information 0.0014308049999999998
label sequence 0.0014130549999999999
current training 0.0013774940000000002
training example 0.0013515880000000001
learning problem 0.001341854
online training 0.001321412
segmentation mod 0.001319959
segmentation candidates 0.0013163060000000002
segmentation score 0.001315135
segmentation agreements 0.001314854
dependency parsing 0.001294779
learning objective 0.001278485
cws models 0.001272281
segmentation phase 0.00127172
tag sequence 0.001260045
ing method 0.001259933
single character 0.001256632
feature 0.00123668
new learning 0.001230554
learning algorithms 0.00119896
gradient method 0.0011981729999999999
final performance 0.00119812
rare features 0.001197477
correlated features 0.001197477
online learning 0.001189752
supervised baseline 0.001189365
unlabeled examples 0.001186381
new words 0.001177043
approach train 0.001152207
recognition performance 0.001148105
unlabeled sentencesdu 0.001144914
learning mechanism 0.001139251
learning cycle 0.001139251
tial models 0.001135349
individual words 0.0011348349999999998
output sentence 0.001132338
unlabeled dataset 0.001115188
test oov 0.001114391
identical words 0.001110068
possible tag 0.001080052
performance scores 0.001070142
label constraints 0.00107003
different strengths 0.001066016
objective function 0.00105965
conditional probability 0.001059367
sequence labeling 0.00104259
supervised case 0.001035001
raw sentences 0.001024456
supervised cws 0.001018683
function com 0.001013401
perceptrons algorithm 0.001009088
training 0.00100696
viterbi algorithm 0.001003149
different behaviors 9.96904E-4
segmentation 9.962E-4
sequence lattice 9.87134E-4
partition function 9.819339999999998E-4
pos tagging 9.798580000000001E-4
score function 9.754E-4
chinese language 9.73001E-4
current sentence 9.610739999999999E-4
statistic information 9.52575E-4
syntactic information 9.51364E-4
