ing constraints 0.002413746
constraints violation 0.002329824
hard constraints 0.002256814
learning model 0.002245415
global constraints 0.002192167
soft constraints 0.0021819170000000002
statistical constraints 0.002156832
unary constraints 0.002126003
constraints dur 0.002121903
constraints valua 0.002121903
feature function 0.00211636
training data 0.002030609
constraints 0.00188437
labeled data 0.0018698439999999998
learning algorithm 0.0017931549999999998
supervised model 0.001785729
constraint violation 0.001778344
unlabeled data 0.001767452
final model 0.0017201590000000002
unsupervised model 0.0016795030000000002
different learning 0.0016742670000000001
current model 0.001673236
data sets 0.001663775
hmm model 0.001637121
data instances 0.0016339929999999998
model parameters 0.001632455
constraint violations 0.001632128
soft constraint 0.001630437
generative model 0.001630261
data size 0.001628505
constraint vio 0.001605154
training algorithm 0.001604509
constraint driven 0.001579693
underlying model 0.0015770250000000001
web data 0.0015764149999999998
weighting data 0.001550563
data pool 0.0015431639999999999
feature functions 0.001482665
training set 0.001456139
local feature 0.001454199
accuracy set 0.0014475970000000001
ing learning 0.001447501
objective function 0.0014342649999999999
distance function 0.0014137269999999998
supervised learning 0.001376564
score function 0.001346702
constraint 0.00133289
model 0.00132729
cost function 0.001299087
labeled training 0.001298193
labeled set 0.001295374
learning procedure 0.001291156
learning algo 0.001290003
boolean function 0.0012856959999999998
other models 0.001276044
machine learning 0.00126864
structured learning 0.001262291
learning process 0.001258924
tance function 0.0012564039999999998
learning case 0.001255041
learning tasks 0.00125413
learning algorithms 0.001253967
learning protocol 0.001200855
supervised training 0.001187918
domain information 0.0011783689999999999
supervised models 0.001177948
learning stage 0.001172887
other inference 0.0011713320000000002
driven learning 0.001164928
large training 0.001151038
new training 0.0011464000000000001
linear models 0.001142938
training examples 0.0011396800000000001
prior knowledge 0.001136078
word similarity 0.001133153
other output 0.001123779
possible output 0.001100344
feature 0.0010979
training sets 0.0010921239999999999
probability distribution 0.001085148
unsupervised models 0.001071722
output variables 0.0010716810000000001
textual word 0.001071095
expressive features 0.001070063
small training 0.001064835
labeled tokens 0.001058139
same labels 0.001057857
other tokens 0.00104596
domain knowledge 0.001037847
training cycles 0.001032027
generative models 0.00102248
transition models 0.001022125
useful information 0.001018992
function 0.00101846
same penalty 0.001013388
final inference 0.001007666
different levels 0.001004395
training samples 0.001004352
tional models 9.96994E-4
inference procedure 9.87828E-4
