output feature 0.002532328
input feature 0.002316144
output features 0.002151238
feature vectors 0.002097712
feature functions 0.002048734
feature types 0.002033365
hypothetical feature 0.001964116
input features 0.001935054
features input 0.001935054
labeling model 0.001817389
probabilistic model 0.001743863
feature 0.00174314
arc features 0.001686208
model parameters 0.0016585830000000001
weighted features 0.001648242
model expectations 0.0016126340000000002
put features 0.001580751
concrete features 0.001578963
direct model 0.001560759
proposed model 0.00153917
learning method 0.001476782
learning approach 0.001409203
features 0.00136205
output label 0.0013618
machine learning 0.001325151
model 0.00132283
optimization algorithm 0.001296207
first factor 0.0012715119999999998
possible output 0.001267961
supervised learning 0.001266132
output distribution 0.001245914
propagation algorithm 0.00124025
probabilistic models 0.001209491
joint learning 0.001200584
many models 0.001175314
output distributions 0.001157052
such tasks 0.00114116
output space 0.001127937
markov models 0.001121101
objective function 0.001088581
output sequences 0.0010873599999999999
likely output 0.0010797
typical models 0.001068009
same input 0.001054866
graphical models 0.001054478
possible label 0.001051385
irregular output 0.001048329
partition function 0.0010442070000000001
input sequence 0.001039836
sequence label 0.001039444
indicator function 0.0010368740000000001
tic models 0.001025299
second factor 0.001024965
tition function 0.001024845
golden output 0.001013362
output spaces 0.001011502
output candidates 0.001011089
annotation data 0.001004678
other predictions 9.97953E-4
such configuration 9.86326E-4
third factor 9.731799999999999E-4
other words 9.68545E-4
sequence labeling 9.61391E-4
learning 9.57024E-4
annotated data 9.54657E-4
notated data 9.393069999999999E-4
different levels 9.36293E-4
algorithm 9.35261E-4
other hand 9.16743E-4
parameters input 9.087570000000001E-4
regular time 9.01715E-4
other decisions 8.96022E-4
structured prediction 8.86115E-4
forward information 8.815399999999999E-4
labeling tasks 8.7885E-4
parameter values 8.78246E-4
information propagation 8.749329999999999E-4
input sequences 8.71176E-4
label sequences 8.70784E-4
confidence information 8.62272E-4
time slices 8.48625E-4
time slice 8.48625E-4
trary time 8.456000000000001E-4
prediction stage 8.42426E-4
rich information 8.26434E-4
conditional random 8.2288E-4
prediction prob 8.164450000000001E-4
same forms 8.11817E-4
multiple sequence 8.1108E-4
label assignments 8.10988E-4
function 8.08111E-4
initial input 8.06638E-4
first term 8.050819999999999E-4
labeling systems 8.042990000000001E-4
ordinary methods 8.025129999999999E-4
information propa 7.940029999999999E-4
ward information 7.92806E-4
final problem 7.905290000000001E-4
information propaga 7.89232E-4
output 7.89188E-4
