response model 0.002367509
translation model 0.0020362419999999997
matching model 0.002005992
good model 0.0019608
retrieval model 0.001931723
ing model 0.001895482
space model 0.0018355589999999998
conversation model 0.001831061
phone model 0.001812899
model parameters 0.0017962759999999999
ranking model 0.0017739569999999998
linear model 0.001768467
line model 0.0017595559999999998
current model 0.0017538579999999998
response models 0.001741906
model map 0.001715837
sponse model 0.0017143
parametric model 0.001705826
trieval model 0.001705826
training data 0.00150121
model 0.00149531
response algorithm 0.001430322
matching feature 0.001423828
response system 0.001387245
matching models 0.001380389
post user 0.001354498
retrieval models 0.00130612
dialogue models 0.001287785
different posts 0.001281814
other words 0.001272054
different entity 0.001264354
data source 0.001263563
parallel data 0.001229899
first pairs 0.0012284409999999998
labeled data 0.001208949
original response 0.0012089240000000001
candidate response 0.001207641
conversation models 0.001205458
natural language 0.001198294
matching features 0.00118199
responses user 0.001177429
automatic response 0.0011599520000000001
ideal data 0.001150655
language conversation 0.001147361
chinese word 0.001135354
first author 0.001134178
baseline models 0.001133632
data cleaning 0.001127766
data with 0.001115858
data extract 0.001115858
original post 0.001106023
response con 0.00110573
word limit 0.00109786
language process 0.001091002
response simpr 0.001082536
response repository 0.001082536
query post 0.001076081
weibo post 0.0010680989999999999
word seg 0.001067913
good responses 0.001057719
features score 0.00105757
language processing 0.0010525530000000001
short text 0.001031612
training set 0.001017229
certain post 0.001015645
short responses 0.001009775
matching function 0.001009647
first day 9.99354E-4
contextual information 9.90556E-4
different choices 9.72671E-4
other users 9.72189E-4
common words 9.701619999999999E-4
different styles 9.69295E-4
long responses 9.52304E-4
original text 9.507910000000001E-4
simple features 9.48074E-4
original responses 9.289540000000001E-4
candidate responses 9.276709999999999E-4
other cases 9.249340000000001E-4
other purposes 9.165410000000001E-4
general setting 9.14808E-4
feature 9.13146E-4
negative pairs 9.12919E-4
other measures 9.01451E-4
matching score 8.96944E-4
sophisticated features 8.96881E-4
labeled responses 8.965749999999999E-4
average responses 8.95637E-4
chinese text 8.930450000000001E-4
mon words 8.864039999999999E-4
long posts 8.85856E-4
score function 8.852269999999999E-4
similar topic 8.85079E-4
response 8.72199E-4
models 8.69707E-4
good performance 8.604999999999999E-4
other parts 8.599780000000001E-4
large candidate 8.53833E-4
english translation 8.46349E-4
quent words 8.44434E-4
