word models 0.002510548
second word 0.002232026
modeling word 0.0022018479999999997
word meanings 0.002188218
word meaning 0.002173535
third word 0.002164359
word mean 0.00215494
word backpack 0.002145776
word person 0.002140134
word chair 0.002135866
lth word 0.002127771
training sentence 0.001898591
model configurations 0.001862872
training sentences 0.001761584
different features 0.0017089380000000001
training corpus 0.0016622920000000001
model 0.0016462
training data 0.0014895239999999999
such training 0.001450719
sentence score 0.001448116
individual words 0.001438158
mapping words 0.001436961
training set 0.001431292
feature vector 0.001415721
training object 0.001407097
sentence length 0.001393502
vidual words 0.0013889969999999999
above sentence 0.001323802
first sentence 0.001321884
size features 0.001313511
particular features 0.00131079
training process 0.00130191
multiple sentences 0.001280251
training sample 0.001241659
incorrect sentence 0.001239751
training sen 0.001229147
sentence tracker 0.001215981
pairs training 0.001213718
corpus score 0.001211817
different representations 0.001207749
image features 0.001205814
whole sentence 0.001194515
several sentences 0.00119385
particular feature 0.001191664
fourth sentence 0.0011857600000000001
discriminative training 0.0011820490000000001
enough training 0.001180519
visual features 0.0011805140000000001
negative training 0.001180181
sentence condi 0.001174618
discrete features 0.001172666
ambiguous training 0.001171289
motion features 0.001169059
tential training 0.001164778
paired training 0.001164593
training samples 0.001163144
ticular features 0.001161595
continuous features 0.001161524
pute features 0.001160616
words 0.00115753
training images 0.001154766
possible state 0.001140728
different methods 0.001133961
individual sentences 0.001106774
feature values 0.0011057
feature vectors 0.00109949
different parts 0.001092237
test data 0.001087918
feature value 0.00108719
such work 0.0010743710000000002
feature vec 0.001059847
whole sentences 0.001057508
different sen 0.001055331
narrated sentences 0.001052594
nth feature 0.001049999
state number 0.001049969
ple sentences 0.0010438700000000001
learning algorithm 0.001018304
different subset 9.8804E-4
log probability 9.86483E-4
different compound 9.85324E-4
different aspects 9.79379E-4
different compo 9.790789999999999E-4
different aspect 9.74839E-4
semantic models 9.65947E-4
sentence 9.63153E-4
transition probability 9.517169999999999E-4
features 9.47316E-4
training 9.35438E-4
state sequence 9.333519999999999E-4
possible track 9.31896E-4
ing data 9.27915E-4
event models 9.26712E-4
above approach 9.2595E-4
test video 9.242670000000001E-4
same video 9.12515E-4
possible paths 9.11649E-4
markov models 9.112829999999999E-4
possible tracks 8.994389999999999E-4
speech csr 8.73007E-4
