word information 0.004088568
possible word 0.003995572
word error 0.0038880470000000004
such word 0.0038332500000000003
word lattice 0.0037224610000000003
baseline word 0.003715152
correct word 0.0037094050000000003
word list 0.003680028
word sequence 0.003652699
original word 0.003638764
word clusters 0.003619705
word lattices 0.003596401
important word 0.0035946610000000003
word lexemes 0.003573693
relative word 0.003570831
oracle word 0.0035643560000000003
word sequences 0.00356297
language model 0.00353592
experiment word 0.0035348180000000003
incorrect word 0.003520703
word clus 0.0035134660000000002
word lexeme 0.003509571
word inser 0.003509571
statistical model 0.002800859
model weight 0.0027713580000000002
probabilistic model 0.002749711
model parameters 0.0027185720000000003
full model 0.0026614480000000003
guage model 0.0026385980000000002
bilistic model 0.002607945
crude model 0.002607945
model 0.00235869
single words 0.002132821
arbitrary words 0.002011676
transcription words 0.002001348
language models 0.001919128
actual words 0.001913433
tual words 0.0019091680000000001
separate words 0.001905839
capitalized words 0.0019027900000000001
words 0.00165266
syntactic information 0.0016382229999999998
language modeling 0.001541782
trigram probability 0.00153242
natural language 0.00145735
inspired language 0.001426165
test data 0.00140409
text corpus 0.001360591
additional probability 0.001347716
attachment probability 0.001339796
probability distributions 0.001289671
syntactic dependencies 0.001288829
joint probability 0.001285258
probability pgram 0.001264923
sentence node 0.001262083
probability dis 0.0012590140000000001
posterior probability 0.0012590140000000001
dependency tree 0.0011999930000000001
statistical models 0.001184067
language 0.00117723
structure grammar 0.001176608
additional information 0.001166794
dictionary information 0.0011569129999999999
test set 0.00114757
single test 0.001146298
parse tree 0.00113467
news data 0.001129798
grammaticality information 0.001121215
transport information 0.001121215
above data 0.001116681
information measure 0.001108692
data figure 0.001103457
only models 0.001089662
likely parse 0.00108819
tual information 0.001084826
statistical parsing 0.001081779
reliable information 0.001080722
textual information 0.001080722
baseline speech 0.001072513
speech recognition 0.001066647
syntactic theories 0.001065499
syntactic properties 0.00106348
sentence boundaries 0.001055729
other com 0.001039971
other hand 0.001039946
structure rules 0.001037781
possible pair 0.001036065
phrase structure 0.0010336920000000001
other children 0.0010206199999999999
parse trees 0.001017633
many ideas 0.001016542
different tags 0.0010150979999999999
flat dependency 0.001013011
parse time 0.001012199
problem structure 0.0010116910000000001
biguous sentence 0.001011159
probability 0.00100924
other events 0.001006669
error rate 0.001005275
error reduction 0.0010027080000000002
