target translation 0.0038110699999999997
word translation 0.00372786
target word 0.00353959
target sentence 0.00309976
target features 0.00299741
source word 0.00291301
target words 0.002665785
target language 0.002663103
source sentence 0.00247318
translation probabilities 0.0024420709999999997
other target 0.002440153
different target 0.002394154
translation pair 0.002394136
target sentences 0.0023640049999999998
data source 0.00234643
candidate target 0.0022132339999999997
target position 0.0021999949999999997
additional target 0.002181477
target positions 0.002153179
entire target 0.002135522
target transla 0.002133681
word level 0.002099999
target fertilities 0.002078225
translation 0.00199967
word cover 0.001995375
other sentence 0.0019171129999999998
sentence pairs 0.001852346
sentence table 0.0018381159999999999
parallel sentence 0.001811552
target 0.0018114
test data 0.001775352
english data 0.001739969
search algorithm 0.0016867380000000001
sentence extraction 0.0016861469999999998
parallel data 0.001684802
feature vector 0.0016839379999999998
sentence pair 0.001682826
feature function 0.001668082
same data 0.0016632230000000001
sentence level 0.001660169
additional sentence 0.001658437
feature set 0.001649518
sentence candidates 0.0016409859999999999
sentence positions 0.0016301389999999999
news data 0.0016221130000000001
baseline data 0.00157755
source position 0.001573415
sentence lengths 0.00155752
additional source 0.001554897
sentence marcu 0.001554736
source document 0.001527783
source positions 0.001526599
spanish data 0.001521152
feature values 0.001498625
fertility features 0.001468577
comparable data 0.001463427
extraction algorithm 0.001460957
gigaword data 0.0014576419999999999
uncovered source 0.001457518
feature functions 0.0014563179999999998
source posi 0.001451383
covered source 0.001451383
lel data 0.001447429
jrc data 0.001429218
parable data 0.001429218
feature value 0.001416231
language pairs 0.001415689
global feature 0.001401076
novel algorithm 0.0013640990000000001
sentence 0.00128836
maxent training 0.001252981
smt system 0.001247208
language pair 0.001246169
training algo 0.001221583
training steps 0.0012160110000000001
weight vector 0.001193581
features 0.00118601
source 0.00118482
probability selection 0.001138275
fertility words 0.001136952
english sentences 0.001130964
glish words 0.001126889
corpus statistics 0.001125615
search space 0.00112151
classification probability 0.001115813
feature 0.00110955
large set 0.0010766249999999999
smt decoding 0.001073955
algorithm 0.00106317
previous work 0.001058131
bleu score 0.001049921
lexicon probabilities 0.001040091
baseline system 0.001035755
matching score 0.001031298
pairs bleu 0.001029052
single hypothesis 0.0010219439999999999
partial score 9.90666E-4
beam search 9.89372E-4
partial hypothesis 9.86108E-4
search states 9.63616E-4
