language model 0.00362352
model size 0.002684194
full model 0.00252852
lossless model 0.0025154549999999998
model representation 0.002513009
model reduction 0.0024314329999999998
language models 0.002410004
combined model 0.002408242
model sizes 0.002377053
guage model 0.002369079
ventional model 0.002362897
model reduc 0.002356946
guagage model 0.002355586
model servers 0.002355586
model server 0.002355586
large language 0.002170636
model 0.00209429
test data 0.002030588
language corpora 0.001975709
training data 0.001968535
target language 0.001955389
lossless language 0.001950395
same translation 0.0018594549999999999
multiple language 0.001854989
randomized language 0.0018526799999999998
translation bleu 0.00183939
data set 0.001837345
language modeling 0.001830351
tributed language 0.0018129309999999998
language mod 0.0018033329999999998
scaling language 0.001796897
ral language 0.00179035
machine translation 0.001765569
lossless data 0.001755695
parallel data 0.001698443
translation experiments 0.0016697790000000001
available data 0.001669148
baseline translation 0.001664707
raw data 0.001657928
data structures 0.0016466480000000001
development data 0.001644742
data sets 0.001629213
data structure 0.001619851
data sources 0.001596341
data track 0.001596341
translation quality 0.001532046
language 0.00152923
byte word 0.001521339
hash function 0.001521097
word ids 0.001506406
word sequence 0.001504703
tinct word 0.001494456
plausible word 0.001493108
word sequences 0.001493108
chine translation 0.001481484
backoff smoothing 0.001418716
random hash 0.0014161809999999999
test test 0.001392116
other experiments 0.001385417
training corpus 0.00136461
error bits 0.0013484500000000002
error rate 0.001335973
same text 0.001329453
perfect hash 0.001307496
large array 0.0012749139999999998
hash functions 0.00127084
other applications 0.001264293
ing corpus 0.00124198
error rates 0.001239455
other parameters 0.001236414
theoretical error 0.0012306040000000002
translation 0.00121506
other hand 0.001214527
test set 0.001198873
other representations 0.001196482
other edges 0.001194614
other areas 0.001194614
variable models 0.001188211
hash func 0.001186774
fingerprint hash 0.0011661459999999998
additional bits 0.001157143
test bits 0.001151933
large set 0.001144221
smoothing scheme 0.001143904
constant space 0.001140922
significant space 0.001140399
further memory 0.001107036
ing size 0.0011012790000000001
further space 0.0010974959999999999
bleu scores 0.001091769
total space 0.0010799170000000001
baseline bleu 0.001073977
unseen text 0.001057811
false positive 0.001056957
single array 0.001056939
corpus statistics 0.001056865
new text 0.001055515
dependent smoothing 0.001046883
backoff weights 0.001038867
space savings 0.001024567
