sentence compression 0.00481731
compression system 0.003498555
standard compression 0.003245721
compression systems 0.003213891
compression rate 0.003199208
compression models 0.003131521
correct compression 0.0030692989999999997
specific compression 0.003026898
man compression 0.003002749
compression rates 0.0029990579999999998
dard compression 0.002986199
compression 0.00270946
sentence length 0.002695632
original sentence 0.002634911
sentence com 0.002615053
corpora sentence 0.0025658110000000003
sentence structure 0.002551884
source sentence 0.002493611
sentence pair 0.002389348
complete sentence 0.002369856
sentence 0.00210785
human compressions 0.0017374320000000001
channel model 0.001732167
regression model 0.0017091390000000001
arc model 0.001692655
other words 0.001681063
other information 0.00167671
training data 0.001505587
syntactic constraints 0.001503731
word editors 0.001499456
new features 0.00147452
human scores 0.001455196
training corpus 0.001411861
model 0.00139835
pos information 0.001392366
level features 0.0013721140000000002
grammatical features 0.001357368
context information 0.001335331
specific features 0.001321878
review data 0.001315132
paired corpus 0.001285214
pendency features 0.0012827760000000002
conjunction features 0.00126799
general news 0.001231781
human ratings 0.001211287
feature set 0.001201807
news articles 0.001173028
syntactic relations 0.0011677459999999999
syntactic head 0.0011632349999999999
new feature 0.001147861
human judges 0.001147233
human judgments 0.001147233
grammatical sentences 0.001135847
compressions gener 0.001117504
target sentences 0.001114559
similar performance 0.001101013
additional sentences 0.001087474
syntactic ana 0.001085423
arc system 0.0010834
structural constraints 0.001068305
presses sentences 0.001046178
discourse level 0.001045332
approach com 0.0010412770000000002
test set 0.001016751
dependency parse 0.001016726
rouge scores 0.0010078069999999999
feature representation 0.001005186
features 0.00100444
training set 0.001000563
feature weights 9.98298E-4
alternative feature 9.81528E-4
evaluation criteria 9.65894E-4
dimensional feature 9.575860000000001E-4
extended feature 9.571740000000001E-4
screen performance 9.52409E-4
parallel corpora 9.50771E-4
pression length 9.48649E-4
words 9.47968E-4
information 9.43615E-4
generalizable feature 9.41335E-4
feature spaces 9.41335E-4
feature encodings 9.41335E-4
corpus 9.35324E-4
target length 9.19422E-4
unsupervised approach 9.181360000000001E-4
original sen 9.06831E-4
gold standard 8.99396E-4
previous approaches 8.984500000000001E-4
future work 8.91275E-4
grammar rules 8.88215E-4
training time 8.77009E-4
constraint approach 8.74258E-4
possible compres 8.71613E-4
compressions 8.53256E-4
programming table 8.5301E-4
different sets 8.51608E-4
pression problem 8.5092E-4
pression rate 8.506150000000001E-4
pos context 8.40467E-4
target com 8.38843E-4
