word error 0.003217798
word set 0.003205439
word sequences 0.002991374
second word 0.0029912600000000004
complete word 0.002964228
word concentration 0.002927306
language model 0.0027910499999999998
speech input 0.002463735
speech recognition 0.002436259
asr system 0.002360388
rover system 0.002112434
system architecture 0.002097233
system outputs 0.002093807
master system 0.002054506
language models 0.002042808
multiple speech 0.002016725
current speech 0.001894549
automatic speech 0.001867863
markov model 0.001865553
parallel speech 0.001856284
speech signal 0.001841213
speech recognizers 0.001837928
speech recog 0.0018244329999999999
heterogeneous model 0.001820524
speech recogniz 0.001816277
speech rec 0.001816277
mainstream speech 0.001816277
system 0.00178428
guage model 0.0017781960000000001
hybrid model 0.001774908
homogeneous model 0.001764196
different grammar 0.0017461669999999999
language software 0.001702262
language processing 0.001692523
unigram language 0.0016642
natural language 0.001644266
own language 0.001603302
phrase resolution 0.0015716830000000002
phrase database 0.0015648160000000001
matching phrase 0.0015583110000000002
input source 0.0015562800000000002
speech 0.0015462
individual words 0.001536344
recognition systems 0.001521387
matching words 0.0015110610000000002
model 0.00149471
ual words 0.001483965
specified words 0.00147845
same input 0.0014515090000000001
recognition results 0.001419852
multiple input 0.00138806
different gram 0.001339689
different automatic 0.001306572
language 0.00129634
same grammar 0.001295232
common recognition 0.001260681
different perspectives 0.001256992
phrase 0.0012524
input sources 0.001251061
recognition engine 0.0012330589999999999
own input 0.001224497
multiple models 0.001216993
input signal 0.001212548
asr systems 0.001207436
words 0.00120515
bad input 0.0011919040000000001
distorted input 0.001187718
recognition accuracy 0.001172394
recognition accu 0.00115933
several systems 0.001137377
unigram grammar 0.0011291180000000001
heterogeneous models 0.001072282
underlying grammar 0.001058673
multiple asr 0.001046633
large grammars 0.0010211069999999998
error rate 9.74918E-4
majority rules 9.746570000000001E-4
software algorithm 9.720989999999999E-4
rules barry 9.64887E-4
individual systems 9.625219999999999E-4
other witnesses 9.52643E-4
alignment module 9.43312E-4
same time 9.38054E-4
input 9.17535E-4
computer systems 9.17157E-4
individual asr 9.07302E-4
error reduction 8.9641E-4
programming algorithm 8.941999999999999E-4
recognition 8.90059E-4
resolution algorithm 8.8546E-4
voting error 8.71538E-4
module scores 8.55524E-4
composite asr 8.45511E-4
early results 8.43686E-4
error rates 8.39666E-4
long way 8.33114E-4
same manner 8.3174E-4
confidence scores 8.10437E-4
unique approach 8.06593E-4
distance scores 7.98992E-4
