model data 0.00280067
training data 0.0022973869999999997
model training 0.002287817
different data 0.002219257
method data 0.002110646
data set 0.002056675
language model 0.0020486339999999997
feature function 0.002037662
other data 0.002023734
large data 0.001899478
test data 0.0018704849999999999
submodular data 0.0018504659999999998
data selection 0.0018253179999999998
single model 0.001823442
translation model 0.001812367
feature set 0.001802705
data sentence 0.0018021959999999998
data subset 0.001795826
modeling data 0.001782266
english data 0.0017781519999999999
nist data 0.001765563
additional data 0.001764087
data sets 0.001762689
natural model 0.001761503
data size 0.0017604679999999998
random data 0.0017590379999999999
small data 0.001740381
good model 0.001733875
parallel data 0.001724548
smt data 0.001720831
data sources 0.001701651
novel model 0.0016801239999999999
original data 0.001663003
data sizes 0.0016609519999999998
german data 0.001653711
based data 0.0016508739999999999
ing feature 0.0016501380000000002
data sub 0.001647509
europarl data 0.001644581
hierarchical model 0.001642476
speech data 0.0016403239999999999
various data 0.0016379699999999999
arabic data 0.001620919
dom data 0.0016204779999999998
reordering model 0.001620299
lingual data 0.001617777
gale data 0.001616422
vant data 0.001616422
guage model 0.001612404
model orders 0.001607454
submodular feature 0.0015964960000000002
feature values 0.001550387
feature weights 0.001547427
training set 0.0015438219999999998
feature length 0.001478474
set features 0.0014701279999999998
nlp feature 0.0014637930000000001
feature weight 0.0014608350000000002
linguistic feature 0.0014402010000000001
feature decay 0.0014258970000000002
model 0.00139555
feature engineering 0.0013836240000000002
raw feature 0.001380162
submodular function 0.001331858
language models 0.001321535
neighborhood function 0.001297161
log function 0.001274909
additional training 0.001251234
experiments function 0.001242317
set functions 0.001235569
domain training 0.0012338409999999998
different translation 0.001230954
complete training 0.00122345
different values 0.0012133740000000001
smt training 0.001207978
weight function 0.001196197
concave function 0.0011886800000000001
function parameters 0.001187404
objective function 0.00118621
function optimization 0.00118502
training samples 0.001172405
such functions 0.001170809
decay function 0.001161259
modular function 0.0011540370000000001
feature 0.00115115
submodular method 0.001150872
original training 0.00115015
large set 0.001145913
linear function 0.001143119
distinct training 0.001140627
target features 0.001135108
root function 0.001134222
cut function 0.001132026
selection method 0.001125724
discriminative training 0.001124871
test set 0.00111692
different concave 0.0011163050000000002
valuation function 0.001115039
available training 0.001112023
saturating function 0.001109846
