translation model 0.00372107
translation system 0.002629844
language model 0.002535128
machine translation 0.002357794
phrase pairs 0.00226962
phrase pair 0.0022161280000000004
large phrase 0.002206684
reordering model 0.0021672230000000002
target phrase 0.002162305
model weights 0.002145695
model weight 0.002130524
phrase table 0.00209408
translation 0.00203494
small phrase 0.0020289080000000003
training data 0.001979489
test data 0.0019746349999999998
model drive 0.001907983
phrase extraction 0.0018970920000000002
word alignment 0.001889151
ble phrase 0.0018619830000000002
stimulus phrase 0.001848093
direct phrase 0.001845264
phrase extrac 0.0017962850000000001
word pairs 0.0017226899999999998
alignment alignment 0.001718102
parallel data 0.001716643
bilingual data 0.001702243
model 0.00168613
training response 0.0016783190000000002
domain response 0.0016406950000000002
conversation data 0.001640147
response pair 0.0016399280000000001
response models 0.0016284580000000002
ing data 0.00162025
phrase 0.00157703
conversational data 0.001574269
human response 0.0015531450000000001
entailment data 0.001537274
commentary data 0.001524924
alignment models 0.0014866789999999999
word alignments 0.0014654639999999999
target language 0.001434273
similar response 0.0014019640000000001
good response 0.0013699790000000001
same language 0.001348087
ibm word 0.001317635
word align 0.001279099
source language 0.00127351
sentence pairs 0.001272404
training corpus 0.001270308
appropriate word 0.001268034
noisy word 0.001267794
identical word 0.001267446
response generation 0.001266102
novel response 0.001257709
response genera 0.001257598
large set 0.001239881
appropriate response 0.001238764
large corpus 0.001222473
propriate response 0.001222426
response generators 0.001220729
response construction 0.001220729
sentence pair 0.001218912
different turkers 0.001208679
phrasal alignment 0.0012034300000000001
large phrases 0.0011733849999999999
human responses 0.0011624970000000002
smt models 0.001143652
smt system 0.0011109280000000002
other words 0.0010994
bilingual pairs 0.001092833
decomposable alignment 0.001079264
random responses 0.001059564
human evaluation 0.0010579
possible responses 0.001050372
exact test 0.001037083
source target 0.001009787
bleu system 0.001007554
parallel corpus 0.001007462
response 0.00100083
smt reordering 9.97117E-4
parallel text 9.93243E-4
same topic 9.77597E-4
feature weights 9.70899E-4
training status 9.676089999999999E-4
pair extraction 9.591600000000001E-4
first work 9.587220000000001E-4
general responses 9.57339E-4
corresponding training 9.571009999999999E-4
media domain 9.53554E-4
conversational training 9.49758E-4
example responses 9.362820000000001E-4
significance test 9.33822E-4
training bitext 9.3171E-4
identical pairs 9.299359999999999E-4
ing responses 9.284320000000001E-4
symmetric pairs 9.24029E-4
training algorithms 9.22176E-4
large portions 9.206080000000001E-4
test statistic 9.20108E-4
