word language 0.00357983
word lattice 0.003278479
recognition word 0.00300978
word lattices 0.002969157
word error 0.002953268
new word 0.002889157
word posterior 0.002884591
indexing word 0.002856151
word recognizer 0.002844715
word pos 0.002839971
word phoneme 0.002816743
single word 0.002803592
phonetic word 0.002795892
word boundary 0.002780662
high word 0.002770775
fom word 0.002769364
regular word 0.002747788
word posteriors 0.002731326
word representations 0.00271999
unified word 0.002718414
merging word 0.0027159849999999998
word transitions 0.002704708
popular word 0.002700556
combining word 0.002697509
combined word 0.002696596
word stringp 0.00269589
associated word 0.002692474
word frag 0.002692474
language model 0.00241117
search query 0.0023559
lattice search 0.002355069
search time 0.002276915
search results 0.002167818
search approach 0.002131307
model probability 0.002119561
search system 0.00208773
search accuracy 0.002030899
ranking search 0.001949139
vocabulary search 0.0019418410000000001
indexing search 0.001932741
phoneme search 0.001893333
linear search 0.001885722
average search 0.001878783
phonetic search 0.001872482
search process 0.001862799
based search 0.001851884
level search 0.001845215
hybrid search 0.001834962
style search 0.00182737
search lookup 0.001822654
search accuracies 0.001803223
unified search 0.001795004
direct search 0.0017907210000000001
search ability 0.001779012
search yields 0.0017737060000000001
search hit 0.001772588
inefficient search 0.001768905
search track 0.001768905
acoustic model 0.001753982
model probabilities 0.001701471
speech recognition 0.0016367600000000001
model weight 0.001621023
language models 0.001608494
lattice lattice 0.001595638
search 0.00155725
speech recognizer 0.001471695
phoneme language 0.001435253
average speech 0.001429173
phonetic language 0.001414402
new words 0.001400613
training data 0.001383049
only speech 0.001364086
conversational speech 0.001358687
trigram language 0.0013499340000000001
speech recog 0.001342173
robust language 0.001333673
inaccurate speech 0.001323117
speech dictation 0.001322716
query phrases 0.001314913
data set 0.00131204
model 0.001312
lvcsr language 0.001310657
lattice audio 0.0012344209999999999
query phrase 0.001232486
short words 0.001211733
known words 0.0012116339999999998
rare words 0.001211579
posterior probability 0.001211492
oov words 0.001210591
information retrieval 0.001208427
inv words 0.0012082969999999999
probability ranking 0.00119945
lattice representation 0.001197975
lattice indexing 0.0011733099999999999
lattice figure 0.001173255
same time 0.00117012
training set 0.001160885
test corpus 0.001160356
query phoneme 0.001134733
phoneme lattice 0.001133902
