model learning 0.00248332
feature set 0.002219554
test model 0.002083609
model representation 0.002029094
trained model 0.0020206490000000002
large feature 0.002004906
feature group 0.001992774
entire model 0.001915791
model complex 0.00189101
model complexities 0.001882469
model complexity 0.001881111
learning algorithm 0.00186059
duce model 0.001856183
compact model 0.001848337
feature space 0.001847538
supervised model 0.00184477
same feature 0.0018414619999999999
tant model 0.00184125
feature weight 0.001787479
feature weights 0.001776395
feature templates 0.00172124
corresponding feature 0.001718898
feature grouping 0.001704609
feature selection 0.001684623
feature string 0.0016744849999999999
age feature 0.001668436
feature strings 0.001667909
feature sets 0.001665922
automatic feature 0.0016504739999999999
tomatic feature 0.001649196
model 0.00159811
learning method 0.0015331210000000001
training data 0.0015034879999999999
feature 0.00140632
learning optimization 0.001356537
loss function 0.0013035479999999999
learning rate 0.001254287
viterbi algorithm 0.001244677
decoding models 0.0012299330000000001
trained models 0.001224977
feasible algorithm 0.001219878
learning framework 0.0012162
base learning 0.001203375
entire learning 0.001202891
learning algo 0.001184944
additional features 0.001175957
test data 0.001151893
online learning 0.0011516
supervised learning 0.00113187
exponential function 0.001120393
parameter space 0.001083397
active features 0.001076651
finite set 0.001069985
related features 0.001068037
accurate models 0.001062899
other words 0.0010609
nite set 0.001058857
compact models 0.001052665
symmetric function 0.001049731
dependency parsing 0.001040902
vector representation 0.00103262
task performance 0.001029389
good results 0.001000544
algorithm 9.7538E-4
method base 9.66076E-4
baseline methods 9.60348E-4
final results 9.53033E-4
many nlp 9.52159E-4
direction method 9.48729E-4
other variables 9.4233E-4
main baseline 9.335230000000001E-4
dimensional vector 9.326790000000001E-4
development data 9.261499999999999E-4
weight values 9.24188E-4
dependency pars 9.168329999999999E-4
tunable parameter 9.167940000000001E-4
ment data 9.113839999999999E-4
standard loss 9.09024E-4
our method 9.073810000000001E-4
form solution 9.04339E-4
quantization method 9.006250000000001E-4
optimization problem 8.989670000000001E-4
exact solution 8.935309999999999E-4
learning 8.8521E-4
gradient methods 8.78344E-4
comparison results 8.76996E-4
admm performance 8.617879999999999E-4
following form 8.53046E-4
dual variables 8.52482E-4
simple form 8.45195E-4
several settings 8.39533E-4
general case 8.392E-4
training 8.37094E-4
discrete values 8.326189999999999E-4
optimization part 8.31328E-4
optimization variables 8.26838E-4
dual decomposition 8.19193E-4
features 8.16451E-4
unique values 8.06043E-4
direct optimization 8.028099999999999E-4
