unlabeled data 0.0033511409999999998
test data 0.003219246
training data 0.003136809
data information 0.0031056599999999997
labeled data 0.00303681
data set 0.003025431
task data 0.0028750449999999997
experiments data 0.00287002
development data 0.002784679
data development 0.002784679
data size 0.002770122
input data 0.002761155
data dev 0.002759015
data increases 0.002707706
unseen data 0.002707088
data coverage 0.0027054519999999997
data abbr 0.002704907
data incor 0.002704907
word table 0.001707343
model parameter 0.001652419
word prediction 0.001648363
probability model 0.00164562
word type 0.0016387119999999999
model structure 0.0016040249999999998
sequence model 0.0015713189999999998
feature vector 0.001564209
good feature 0.0015367389999999999
supervised learning 0.0015319069999999999
markov model 0.0015194549999999999
model ssl 0.001519182
test test 0.001508972
conditional model 0.0015037219999999999
discriminative model 0.0014838689999999999
training algorithm 0.001479649
simple model 0.0014737329999999999
learning methods 0.001466413
hybrid model 0.001459029
graphical model 0.0014561049999999999
scalable model 0.001430993
sophisticated model 0.001423833
native model 0.0014162939999999998
common feature 0.001388294
feature templates 0.0013717109999999999
feature repre 0.00137082
feature template 0.0013669899999999998
feature tem 0.0013641999999999999
feature design 0.001362075
deep feature 0.0013610369999999998
feature engineering 0.0013610369999999998
test set 0.001315157
machine learning 0.001312627
same training 0.0012714129999999999
large test 0.0012587599999999998
discriminative learning 0.001255909
labeled training 0.001244099
same graph 0.001239087
learning applications 0.001213231
supervised crf 0.001212935
estimation algorithm 0.0012034559999999999
test sets 0.00119499
learning applica 0.001188195
model 0.00117535
system test 0.001162937
additional features 0.001149279
optimization algorithm 0.001139368
crf results 0.001136244
large performance 0.001122279
feature 0.00111978
above test 0.001113132
crf parameter 0.001105487
same way 0.001081122
development test 0.001074405
test development 0.001074405
graph structure 0.001068398
performance tagger 0.001056366
segmentation problems 0.001054623
test dev 0.001048741
performance improvements 0.001041882
significant performance 0.001003231
test collections 9.97546E-4
test collec 9.96213E-4
test col 9.951509999999999E-4
ssl method 9.82299E-4
discriminative training 9.80568E-4
estimation approach 9.72397E-4
auxiliary problems 9.681819999999999E-4
same procedure 9.58473E-4
joint probability 9.58355E-4
ner performance 9.54831E-4
same form 9.51606E-4
tagging results 9.507469999999999E-4
left graph 9.4988E-4
generative models 9.47741E-4
learning 9.4739E-4
auxiliary problem 9.45857E-4
high performance 9.44478E-4
target problem 9.42543E-4
pos tagging 9.312599999999999E-4
efficient training 9.27577E-4
right graph 9.253530000000001E-4
