chunk language 0.002171049
system training 0.001924401
chunk parser 0.00192142
speech recognition 0.001887966
training data 0.001863731
parsing system 0.0018362
system evaluation 0.0018153219999999999
word hypotheses 0.00174645
word error 0.001745164
data set 0.001727415
set words 0.001724635
language model 0.00170663
same data 0.0016980160000000001
reranking system 0.001687585
system performance 0.001685672
chunk score 0.0016777440000000001
ing system 0.001658768
score words 0.0016477940000000002
words score 0.0016477940000000002
chunk structure 0.0016470410000000001
word accuracy 0.001635191
several words 0.001630254
chunk sequence 0.0016294270000000001
multiple word 0.00161143
correct chunk 0.001608656
word gain 0.001585019
system architecture 0.0015836519999999998
chunk phrases 0.001579508
chunk accuracy 0.001577821
phrase chunk 0.00157577
wer wer 0.00156665
language models 0.0015620489999999998
global system 0.0015567659999999998
word errors 0.001547518
third word 0.0015388749999999999
word stream 0.0015372609999999998
word sequences 0.001531544
speech recognizer 0.0015277609999999999
based chunk 0.001527109
frequent word 0.001521587
fect word 0.001514011
word repeti 0.001513187
ithe word 0.001513187
data analysis 0.001511427
chunk coverage 0.001507171
chunk representations 0.001502138
system evalua 0.001491315
system runtime 0.001491315
following data 0.001483455
data sets 0.001482214
speech tagger 0.001478821
complex chunk 0.00147601
chunk gram 0.001474735
chunk sequences 0.001474174
duce chunk 0.001470453
chunk lan 0.001465899
chunk representation 0.0014644900000000001
conjunction chunk 0.0014591460000000001
chunk cov 0.00145597
tli chunk 0.00145597
data the 0.001443513
the data 0.001443513
consortium data 0.0014350580000000001
data consortium 0.0014350580000000001
frequent words 0.0014342670000000001
board data 0.001430354
hesitation words 0.001428374
unparsable words 0.001428374
noise words 0.001428374
different pos 0.001426508
different hypotheses 0.001414341
speech recog 0.001394598
speech face 0.001384928
conversational speech 0.0013817389999999999
speech sequences 0.0013795139999999999
continuous speech 0.001377433
speech recognizers 0.001376635
spontaneous speech 0.001375168
janus speech 0.001363281
speech dysfluencies 0.0013617389999999998
natural language 0.001355369
language mod 0.001302188
alternative language 0.001266934
system 0.00123954
training set 0.001233406
performance wer 0.001229457
model scores 0.001229042
model score 0.001213325
chunk 0.00120604
first hypothesis 0.001192527
different est 0.00118373
different judgement 0.001183
words 0.00117609
features 0.00114312
optimal wer 0.001137131
syntactic structures 0.00112511
significant wer 0.001121212
average wer 0.001115135
speech 0.00111138
wer gain 0.001104934
