language model 0.0039600699999999996
language models 0.00281017
model probabilities 0.002772854
baseline model 0.002747854
standard model 0.0025891549999999997
line model 0.0025039999999999997
guage model 0.002471797
plain model 0.002457409
such language 0.002385433
baseline language 0.002309284
large language 0.002276187
target language 0.002211021
model 0.00219932
filter language 0.00214236
gram language 0.002081792
conventional language 0.002062246
language modelling 0.002023212
language modelsmay 0.002018647
randomised data 0.0018814029999999998
lossless data 0.001785675
language 0.00176075
data structure 0.0017519529999999999
such models 0.001674103
associative data 0.0016566749999999998
allel data 0.001641451
baseline models 0.001597954
hash values 0.001589698
other smoothing 0.0015167750000000002
hash functions 0.0014557860000000001
translation performance 0.0014306689999999999
training corpus 0.001411001
training set 0.001387955
translation experiments 0.001380466
srilm models 0.001377475
hash map 0.001373452
machine translation 0.001333337
training frequency 0.001326644
error experiments 0.001319295
memory bits 0.001314391
trigram models 0.001309213
test time 0.0012391910000000002
other schemes 0.001236409
bit array 0.001232087
small test 0.0012285199999999999
error rate 0.001227785
test set 0.001224456
error analysis 0.001220975
same corpus 0.001201128
training scheme 0.001192334
smoothing schemes 0.001179996
bleu scores 0.001165754
test frequency 0.001163145
function character 0.001149677
different approach 0.0011376110000000002
actual error 0.001135703
other items 0.001134252
ing sentence 0.001128904
mer training 0.001128327
rent word 0.001128278
context count 0.001125172
overestimation error 0.001111141
previous value 0.001101787
feature functions 0.00109762
boolean function 0.0010963029999999999
target words 0.001090145
negative probability 0.00108696
effective error 0.001086905
base bleu 0.00108493
underlying error 0.001084514
memory requirements 0.0010820040000000001
memory savings 0.001078132
first set 0.001074846
distinct set 0.001074224
frequency information 0.001073013
minimum error 0.001068062
error guarantees 0.001063247
small number 0.001060741
conditional probability 0.001057672
other hand 0.001056915
smoothing scheme 0.001054601
significant space 0.0010541209999999999
parallel text 0.00105226
models 0.00104942
other components 0.001047435
additional proxy 0.001036377
bleu score 0.00103447
membership probability 0.001026176
same corpora 0.001020077
suffix counts 0.001013851
corresponding value 0.001013119
quantifiable probability 0.00101085
small cache 0.001001948
proxy space 9.95714E-4
false positive 9.95167E-4
additional savings 9.94454E-4
test sentences 9.93247E-4
suffix count 9.92604E-4
expected value 9.78756E-4
such bounds 9.78029E-4
test item 9.77076E-4
