language model 0.00307845
translation model 0.0028135260000000002
model first 0.0024858470000000002
word length 0.002411581
typing model 0.00240579
new word 0.002366898
user model 0.00230041
model table 0.002297913
next word 0.0022941430000000002
topic model 0.002279953
trigram model 0.0022673850000000002
correct word 0.002244978
short word 0.002242572
entire word 0.002228595
gram model 0.0022106030000000002
current word 0.002201475
intended word 0.002200704
mixture model 0.002200224
selection model 0.002197396
word predictions 0.002190159
word choices 0.002188497
word rankw 0.002188497
model size 0.002180412
stimulus model 0.002179616
comparable model 0.0021770210000000003
model complex 0.0021661930000000003
election model 0.002150213
channel model 0.002127546
ture model 0.002120139
language models 0.0019566080000000003
model 0.00189476
response language 0.001859917
target language 0.001654357
previous words 0.001634314
language modeling 0.001593617
training data 0.0015744420000000001
candidate words 0.001522585
source language 0.001521425
top language 0.001502057
gram language 0.001499533
content words 0.001473379
preceding words 0.001462483
vious words 0.001439945
representative words 0.001436929
proceeding words 0.001432339
generic language 0.001431534
language prediction 0.001421745
test data 0.001415924
language predictability 0.001411183
other models 0.001405329
language usage 0.001402608
translation task 0.0013969569999999999
machine translation 0.00134106
translation table 0.001321919
approach models 0.0012992820000000001
news data 0.001284029
parallel data 0.001279406
data source 0.001246943
words 0.0012121
comment data 0.001205207
several models 0.001201239
development data 0.001194341
language 0.00118369
data sparseness 0.001163669
different task 0.0011569459999999998
comments data 0.0011489059999999999
trigram models 0.0011455430000000002
target text 0.001124206
topic distribution 0.001123285
response length 0.001118548
different settings 0.001102663
different input 0.001095088
mixture models 0.0010783820000000001
text input 0.001069872
corpus news 0.0010395439999999999
different parameter 0.001031592
response message 0.001023902
other set 0.001019588
other topic 0.001017604
prediction models 0.001010973
response completion 0.001010464
different phrases 0.001010418
additional information 0.00100595
other methods 0.001003805
different vocabulary 0.001003381
response generation 9.98661E-4
source text 9.912739999999999E-4
direct response 9.91241E-4
english news 9.89902E-4
text completion 9.877759999999999E-4
different values 9.81563E-4
different users 9.81134E-4
conditional probability 9.77463E-4
other translations 9.71733E-4
such estimates 9.67384E-4
different types 9.630699999999999E-4
stimulus response 9.61083E-4
zipf distribution 9.58569E-4
new information 9.500560000000001E-4
proper training 9.486760000000001E-4
