such features 0.00213697
language model 0.002081173
similar feature 0.001988012
new feature 0.001925455
hmm model 0.00191706
feature representation 0.0019008699999999998
model selection 0.001833632
third feature 0.0018317099999999998
feature type 0.0018293019999999999
constant feature 0.001824725
feature map 0.001822209
modeling features 0.001820208
feature rescaling 0.001812923
suitable feature 0.00178466
binary features 0.0017824100000000001
edge features 0.001781192
node features 0.001771849
syntactic features 0.0017435570000000002
features category 0.001725517
level features 0.00172502
design features 0.001720625
surface features 0.001690102
bow features 0.001681262
appropriate features 0.001679496
training algorithm 0.0016765880000000001
training method 0.001595623
same algorithm 0.001589146
feature 0.00153076
model 0.00150746
learning method 0.001493504
features 0.00142606
sequence models 0.0014198190000000001
first word 0.001405957
same set 0.0013217480000000002
loss function 0.001315985
margin training 0.0012958660000000001
other learning 0.0012806229999999998
programming algorithm 0.001263318
constant function 0.001250361
same paragraph 0.001247332
cost function 0.00124263
algorithm input 0.001242054
other sequence 0.001232374
such text 0.001219343
sophisticated function 0.001215279
criminant function 0.001210701
viterbi algorithm 0.001199955
efficient algorithm 0.0011833920000000001
generation algorithm 0.001176285
previous models 0.001158196
see algorithm 0.0011497
training error 0.001146294
competitive algorithm 0.001139417
markov models 0.001132328
word overlap 0.001130312
art method 0.0011267529999999999
sequence label 0.0011122530000000001
linear time 0.001105412
method pub 0.001097636
ing information 0.001096077
english data 0.0010896039999999999
discriminative learning 0.001083461
time inference 0.0010808179999999999
training instances 0.001076221
output text 0.00105199
modeling sequence 0.001038951
ing problem 0.001037751
graphical models 0.001036277
test set 0.001034896
machine learning 0.001027585
german data 0.001024258
learning algorithms 0.001023704
paragraph structure 0.001007787
ger data 0.001005749
development data 0.00100517
sequence length 0.00100324
large margin 0.0010022569999999999
tion sequence 0.001000587
sequence segments 9.97138E-4
binary classification 9.96417E-4
classification task 9.90553E-4
sparse data 9.80896E-4
same author 9.76364E-4
learning australian 9.723399999999999E-4
margin constraint 9.7172E-4
same tech 9.69579E-4
language modeling 9.67861E-4
same book 9.62854E-4
current set 9.571810000000001E-4
function 9.56396E-4
first corpus 9.56251E-4
test results 9.48785E-4
sequence segmentation 9.4867E-4
sequence prediction 9.480280000000001E-4
first segment 9.47446E-4
optimization problem 9.42514E-4
sequence seg 9.39815E-4
maximum margin 9.37029E-4
maximal sequence 9.34094E-4
short sequence 9.32628E-4
