sentence compression 0.0036179999999999997
original sentence 0.0026052049999999998
sentence length 0.002547379
short sentence 0.002470212
sentence pairs 0.0024280449999999998
long sentence 0.002408184
grammatical sentence 0.0023655359999999997
goal sentence 0.002353395
input sentence 0.002352393
sentence com 0.002332211
complex sentence 0.0023255809999999997
nal sentence 0.00227713
plex sentence 0.002273117
probabilistic model 0.002177402
channel model 0.002167979
compression method 0.002162572
source model 0.002162078
entropy model 0.002123298
compression problem 0.002069601
tropy model 0.002057301
sentence 0.00202829
compression methods 0.0019639180000000003
compression rate 0.001947865
average compression 0.001899361
compression forest 0.001880652
compression ratio 0.001844784
compression ratios 0.001840868
compression knight 0.0018362300000000001
model 0.00181244
original sentences 0.0016829750000000002
same sentences 0.0016715340000000001
many tree 0.0016198879999999999
compression 0.00158971
tree node 0.001583417
test sentences 0.0015703400000000001
good sentences 0.0015660090000000002
short sentences 0.001547982
parse tree 0.0015409689999999999
training corpus 0.001526347
training data 0.001512592
few sentences 0.00151248
short tree 0.001487652
long sentences 0.001485954
tree structure 0.001477385
new tree 0.0014693789999999998
input sentences 0.0014301630000000001
other information 0.001426114
different training 0.00141249
complex sentences 0.0014033510000000002
tree pair 0.0013909389999999999
tree structures 0.001369215
complicated sentences 0.001362013
nal sentences 0.0013549
test corpus 0.001335634
word recall 0.001335459
inal tree 0.00129181
different length 0.0012765860000000001
parallel corpus 0.001276291
other problem 0.001258191
language models 0.001248759
evaluation score 0.001239383
many features 0.001230803
feature function 0.0012064839999999999
generation rules 0.00120013
human output 0.001199831
corpus bleu 0.0011904489999999999
learning features 0.00119027
corpus size 0.001184926
syntactic trees 0.00118237
length information 0.001166903
velopment corpus 0.001164446
other methods 0.001152508
important words 0.001144659
evaluation method 0.001142042
human summary 0.001139868
original rule 0.001134531
cfg rules 0.0011283040000000001
corpus sizes 0.001120348
davis corpus 0.0011195929999999999
sentences 0.00110606
negative words 0.001104811
other hand 0.0010966370000000001
common words 0.00108428
original parse 0.001072154
probability distribution 0.0010680870000000001
other characteristics 0.001067047
other parts 0.001063998
method grammar 0.001061377
automatic summarization 0.001045945
same problem 0.001045365
output length 0.0010446399999999999
system summary 0.001042078
other applications 0.00104122
special rules 0.001038184
different lengths 0.001036984
specific features 0.0010221689999999999
high score 0.001006688
different aver 0.001005688
necessary words 9.966649999999999E-4
extraneous words 9.96661E-4
