language model 0.0029346
test data 0.002897823
training data 0.0028803329999999997
parallel data 0.002596238
randomised data 0.002481785
ing data 0.00246365
new data 0.002462945
stream data 0.002398739
data stream 0.002398739
blog data 0.00232332
translation results 0.002321314
other model 0.002315516
lossless data 0.002311811
data structure 0.002280277
data structures 0.002259136
recent data 0.002258787
input data 0.002249212
much data 0.0022484519999999997
model error 0.002240819
data streams 0.002230012
monolingual data 0.0022261449999999997
novel data 0.00222382
data date 0.0022139029999999997
language models 0.002202702
textual data 0.002196736
data insertion 0.002188147
translation system 0.002126701
memory translation 0.002096899
translation performance 0.0020875760000000003
machine translation 0.002081029
baseline translation 0.0020675900000000002
current translation 0.0020576730000000003
translation experiments 0.002051817
target language 0.002005529
randomised language 0.001978965
translation task 0.001978646
online model 0.001973193
previous translation 0.00195635
translation quality 0.001938936
translation streaming 0.0019318270000000001
translation setting 0.0019264100000000001
translation point 0.001916311
language stream 0.001895919
chine translation 0.001895427
translation points 0.0018952740000000002
translation literature 0.0018925780000000002
translation perfor 0.0018895910000000001
translation scenario 0.0018895910000000001
exact model 0.0018824129999999999
different test 0.001800359
exact language 0.001792293
gram model 0.0017903259999999999
model parameters 0.001790008
guage model 0.001784426
domised model 0.001775231
adaptive language 0.001699693
language mod 0.001695222
domised language 0.001685111
big language 0.001683941
same training 0.0016651729999999998
translation 0.00162771
rate training 0.001543301
model 0.00151236
new training 0.001493158
small space 0.001461743
same hash 0.001435187
test set 0.001425514
language 0.00142224
training set 0.001408024
hash function 0.001390681
parallel corpus 0.001379306
backoff probability 0.00134549
last test 0.001341164
subset test 0.001337459
error rate 0.0013164869999999999
same target 0.001293189
test sets 0.001268421
test times 0.001266976
training purposes 0.001263454
test material 0.001262196
test point 0.001261364
target text 0.0012589519999999998
monolingual training 0.001256358
large corpus 0.0012521910000000002
fixed training 0.001248339
training material 0.001244706
online models 0.001241295
significant space 0.001241177
models performance 0.001240328
test points 0.0012403269999999998
test dates 0.001236477
ing results 0.001232194
ing approach 0.001207331
same domain 0.001199354
other events 0.001193946
high probability 0.001192555
random hash 0.001183482
hash functions 0.001178476
space frequencies 0.001171718
space savings 0.00116421
