language model 0.00381424
language models 0.0031383500000000003
test data 0.002616752
new language 0.002588406
training data 0.002580243
tree language 0.00256104
language system 0.002556404
data results 0.0025499809999999998
gram language 0.002507057
model probability 0.002487001
language processing 0.002450125
language modeling 0.0024470480000000003
based language 0.00243545
natural language 0.002392206
basic language 0.0023908500000000004
random data 0.002387822
language mod 0.0023704810000000002
forest language 0.0023576070000000003
model smoothing 0.00235116
type language 0.002342045
data size 0.002302139
data events 0.002238254
model probabilities 0.002225418
data ppl 0.00219452
data sample 0.002169294
data structure 0.002166977
data statistics 0.002155874
data likelihood 0.002155502
efficient data 0.00215343
heldout data 0.002153416
unseen data 0.002144395
data sparseness 0.0021416869999999998
gram model 0.002139657
data fragmentation 0.002112031
language 0.00209082
guage model 0.002068331
final model 0.002025519
trigram model 0.002009862
unseen model 0.002006625
model evalution 0.001976415
pact model 0.001968349
model 0.00172342
word vocabulary 0.0016450009999999999
word error 0.0016303569999999998
word string 0.001587574
word sequence 0.0015777539999999998
network models 0.0015049310000000002
next word 0.0015031159999999999
speech recognition 0.0015011760000000001
segment word 0.001495696
frequency words 0.001479259
vocabulary speech 0.001478691
gram models 0.0014637670000000001
guage models 0.0013924410000000001
automatic speech 0.0013867390000000001
same training 0.00135535
unknown words 0.001350997
natural speech 0.0013239760000000001
new approach 0.0013206709999999998
test text 0.00127812
test set 0.001271963
lary speech 0.0012713310000000001
speech recog 0.001270267
training corpus 0.00124113
error test 0.0011970190000000001
different task 0.0011905800000000001
machine translation 0.001186566
same time 0.001179866
modeling approach 0.001179313
backoff smoothing 0.001162692
order probability 0.001151749
test events 0.0011326259999999999
different values 0.001108661
words 0.00110006
probability distribution 0.0010928
similar results 0.001090729
smoothing method 0.001082599
syntactic information 0.0010812999999999999
greedy approach 0.001070077
eling approach 0.001070077
trial probability 0.0010672260000000001
large vocabulary 0.001065242
probability estimation 0.001053769
heldout test 0.001047788
different tasks 0.001047611
models 0.00104753
other tasks 0.001041499
unseen test 0.001038767
pairs sentence 0.001031743
test setup 0.0010300399999999999
various smoothing 0.001027553
same order 0.001024465
new random 0.001024218
success probability 0.0010236450000000001
speech 0.00102259
ppl results 0.001022121
aggregated probability 0.0010198240000000001
equal probability 0.00101589
following form 9.80358E-4
sentence segment 9.80181E-4
