language model 0.00346212
repair word 0.0033968699999999998
sentence word 0.003154928
source word 0.0029768670000000002
same word 0.002903239
ing word 0.002754066
source model 0.002734737
word string 0.002685071
reparandum word 0.0026672090000000002
word order 0.002649321
word classifier 0.002646416
repair words 0.00263958
model tag 0.0026257119999999997
word pairs 0.002609612
model performance 0.0025839219999999998
word window 0.002573469
word flight 0.002566073
such model 0.002559271
word basis 0.002556718
temporal word 0.002556114
word bigram 0.0025545480000000002
word tomorrow 0.002546848
preceding word 0.00254408
dum word 0.002542132
channel model 0.002414483
generative model 0.002413961
trigram model 0.002398224
probabilistic model 0.002375034
noisy model 0.0023323709999999997
model design 0.002313767
bigram model 0.002312418
tic model 0.002309896
guage model 0.002302896
model incremen 0.002292034
combined model 0.002287912
regnum model 0.002285387
actual model 0.002283374
nel model 0.002282577
model charniak 0.002279774
language models 0.0022296
source words 0.002219577
model 0.00205859
ing words 0.001996776
reparandum words 0.001909919
repair speech 0.0019090679999999999
similar words 0.0018987029999999998
simple language 0.001885696
interregnum words 0.001841892
state language 0.0018275280000000001
words costs 0.001826359
repair data 0.001803639
partial words 0.0017891629999999999
parser language 0.001788432
preceding words 0.00178679
cal words 0.0017864559999999999
tial words 0.0017749629999999998
words boston 0.001768731
words shriberg 0.001764561
probabilistic language 0.001719974
gram language 0.001694958
bigram language 0.001657358
language terms 0.001646831
ing repair 0.001549496
words 0.00154343
source sentence 0.001530355
repair detection 0.001526065
training data 0.001506054
machine translation 0.001495447
repair string 0.001480501
denver repair 0.001468434
high repair 0.001459217
repair channel 0.0014520429999999999
training corpus 0.00140367
language 0.00140353
corresponding repair 0.001376626
repair substring 0.001366903
repair strings 0.0013655289999999999
last repair 0.001356398
typical repair 0.00135263
repair type 0.00134681
repair odds 0.001318528
training sentences 0.001268068
many speech 0.001248258
speech repairs 0.001209572
main training 0.001185454
test data 0.001184784
statistical speech 0.001178187
probability distribution 0.001168464
source sentences 0.00114565
probabilistic models 0.001142514
probabilistic speech 0.001129362
high probability 0.001122912
translation 0.00112165
underlying sentence 0.001111126
speech recognition 0.001099339
joint models 0.001097771
repair 0.00109615
whole sentence 0.001092046
sentence marker 0.001091377
unigram models 0.001080093
