language model 0.001874625
training data 0.001746628
language models 0.001596209
rank language 0.001581407
corpus data 0.0015465309999999999
order model 0.001489776
parallel data 0.001423091
monolingual data 0.0013560909999999998
data sparsity 0.0013289649999999999
mentary data 0.001318168
translation task 0.001306611
model changes 0.001288497
general algorithm 0.0012601280000000001
unigram model 0.0012568739999999998
word sequence 0.001242701
model outputs 0.001231467
word classes 0.001187751
class language 0.001183407
neural language 0.0011775940000000001
rank approach 0.00117646
next word 0.001173145
language modeling 0.001167894
unseen word 0.001157702
machine translation 0.001154356
training size 0.001153865
word sequences 0.001152786
word types 0.001149624
legitimate word 0.001141738
rank approximation 0.001092648
ral language 0.0010871000000000001
full rank 0.0010782440000000001
rank dataset 0.001077248
elegant algorithm 0.001075568
test set 0.001069785
training time 0.001064716
phrase translation 0.001060369
small training 0.001056749
low rank 0.001045776
common words 0.001042125
rank matrix 0.001040021
trigram models 0.001032453
rank approaches 0.001022207
rank bigram 0.0010204810000000002
rank terms 0.001020229
rank trigram 0.001017651
test time 0.001013194
model 0.00100742
probability vector 9.98156E-4
interpolated models 9.94561E-4
rank representation 9.874200000000001E-4
rank approximations 9.812520000000002E-4
other methods 9.81025E-4
maximum rank 9.80091E-4
rank matrices 9.78789E-4
distinct words 9.71858E-4
rank quantity 9.69763E-4
rank structures 9.69726E-4
family models 9.69313E-4
training cor 9.6528E-4
training sets 9.64962E-4
gram models 9.64593E-4
rank quantities 9.64087E-4
task table 9.59594E-4
unique words 9.51676E-4
cial rank 9.497360000000001E-4
single parameter 9.496929999999999E-4
rank bigrams 9.40927E-4
probability estimates 9.40741E-4
rank tensors 9.38562E-4
plre training 9.36975E-4
rank ensembles 9.368870000000001E-4
rank matri 9.362960000000001E-4
rank approxi 9.35614E-4
discount function 9.354210000000001E-4
rank tri 9.340290000000001E-4
rank trigrams 9.340290000000001E-4
rank struc 9.340290000000001E-4
wmt translation 9.23247E-4
training times 9.20888E-4
first review 9.154829999999999E-4
test sets 9.1344E-4
first row 9.076169999999999E-4
count matrix 8.9657E-4
probability table 8.94623E-4
bigram count 8.7703E-4
bigram counts 8.75282E-4
fast training 8.742600000000001E-4
original test 8.73734E-4
smoothing methods 8.71766E-4
other applications 8.6736E-4
language 8.67205E-4
evaluation time 8.56678E-4
conditional probability 8.56372E-4
order distribution 8.56142E-4
large number 8.48618E-4
other smooth 8.48243E-4
algorithm 8.47995E-4
original counts 8.45831E-4
test complexity 8.44538E-4
other hand 8.41477E-4
