language model 0.001998581
model weighted 0.001769017
statistical model 0.001546449
such states 0.00154255
model transducer 0.0015115670000000001
transducer model 0.0015115670000000001
statistical language 0.00150875
language models 0.0015084260000000002
language processing 0.001386208
bigram model 0.0013756130000000001
natural language 0.0013520250000000002
probabilistic model 0.0013418170000000001
backoff model 0.001333687
joint model 0.0013233350000000001
model composition 0.0013123050000000002
input string 0.001309293
backoff language 0.001295988
training corpus 0.0012878
standard speech 0.0012857950000000002
trigram language 0.0012745730000000002
weighted transducer 0.001244304
tical model 0.00123887
other unit 0.001204519
tical language 0.001201171
stochastic language 0.001199933
final state 0.001197294
initial state 0.001191796
training data 0.001164508
such tests 0.001157948
simple algorithm 0.0011419210000000002
backoff state 0.001128465
weighted automaton 0.001125283
other methods 0.001124199
following corpus 0.001107832
same input 0.001099087
new cost 0.001094697
new unit 0.0010887829999999999
output string 0.001071696
context cost 0.001066303
statistical models 0.0010562940000000002
statistical unit 0.0010483390000000001
finite input 0.001038815
phonetic context 0.001036641
feature cost 0.001035114
coaccessible state 0.001032959
origin state 0.001029951
weighted difference 0.001029582
weighted automata 0.001021908
model 0.00101814
input label 0.001015489
other components 0.001006932
weighted transducers 0.001006615
standard error 0.0010063890000000001
unit sequence 0.001004453
final states 0.001004273
composition algorithm 0.001003815
initial states 9.98775E-4
modeling approach 9.89852E-4
large corpus 9.83869E-4
product speech 9.8375E-4
string semiring 9.823710000000001E-4
language 9.80441E-4
classical algorithm 9.79244E-4
new system 9.75767E-4
statistical methods 9.680190000000001E-4
transducer mapping 9.676209999999999E-4
general approach 9.65746E-4
speech recognition 9.65189E-4
training cor 9.49819E-4
target cost 9.49049E-4
input halfphone 9.381249999999999E-4
final weight 9.37888E-4
grammar cost 9.375329999999999E-4
backoff states 9.35444E-4
statistical system 9.353230000000001E-4
initial weight 9.3239E-4
training utterance 9.31355E-4
new systems 9.31343E-4
other range 9.312810000000001E-4
appropriate training 9.30081E-4
extended input 9.2527E-4
standard beam 9.20021E-4
standard research 9.176500000000001E-4
statistical modeling 9.16781E-4
acoustic cost 9.16187E-4
finite set 9.15647E-4
other elements 9.132630000000001E-4
probability distribution 9.08199E-4
string potential 9.04692E-4
input alphabet 9.01338E-4
conditional probability 9.003450000000001E-4
input labels 8.98878E-4
standard mean 8.960610000000001E-4
speech synthesis 8.871040000000001E-4
selection algorithms 8.84962E-4
standard shrinking 8.83556E-4
standard fashion 8.826250000000001E-4
learning algorithms 8.81345E-4
string potentials 8.73266E-4
put string 8.68157E-4
