visual attribute 0.001741747
type attribute 0.0017196569999999999
color attribute 0.0015700409999999999
size attribute 0.0015392499999999998
visual object 0.001490403
other words 0.001439658
attribute descriptors 0.0014191199999999998
cation attribute 0.001399682
other work 0.001370792
target object 0.001343786
spatial relations 0.001298656
small object 0.001298616
object recognition 0.001292878
visual objects 0.001290183
different data 0.001255054
different cost 0.0012060740000000001
attribute 0.00119499
visual features 0.001190043
other nodes 0.001183037
other approaches 0.001182203
different graph 0.001162398
binary relations 0.001158157
gray object 0.001147838
target objects 0.001143566
collaborative model 0.001140112
ing information 0.00111751
relative attributes 0.001102787
tial relations 0.0011020049999999999
visual scene 0.001096771
other hand 0.001095216
several objects 0.001086214
ing user 0.001085357
similar objects 0.00107931
human models 0.001078991
different generation 0.001066699
visual feature 0.0010646190000000002
type information 0.001059165
situated dialogue 0.00103129
different strategies 0.0010278330000000001
different scenes 0.001024826
search algorithm 9.99624E-4
human descriptions 9.95171E-4
previous work 9.87215E-4
cost function 9.85239E-4
different types 9.72713E-4
robot dialogue 9.68121E-4
large set 9.61922E-4
category words 9.60285E-4
reg performance 9.599420000000001E-4
sual objects 9.58169E-4
absolute performance 9.50058E-4
uated dialogue 9.472759999999999E-4
object 9.43646E-4
specific words 9.35575E-4
different genera 9.311510000000001E-4
different means 9.26803E-4
numerical features 9.26107E-4
current approach 9.23003E-4
human strategies 9.13471E-4
type category 9.12981E-4
many recognition 8.97113E-4
scene figure 8.90435E-4
relative ones 8.8975E-4
natural language 8.85037E-4
domain information 8.84953E-4
relations 8.80312E-4
absolute attributes 8.76713E-4
big performance 8.758330000000001E-4
many processing 8.66255E-4
many approaches 8.62397E-4
model 8.57735E-4
reliable features 8.574629999999999E-4
such discrepancies 8.56766E-4
original approach 8.555E-4
visual scenes 8.49229E-4
scene hypergraph 8.483540000000001E-4
human partners 8.440520000000001E-4
perceptual basis 8.423290000000001E-4
numerical value 8.38044E-4
visual world 8.35595E-4
semantic grounding 8.337490000000001E-4
human robot 8.33134E-4
many errors 8.307599999999999E-4
relative preferred 8.30608E-4
performance gap 8.296550000000001E-4
training data 8.28699E-4
human part 8.28559E-4
referential descriptions 8.24401E-4
human partner 8.2429E-4
class type 8.22917E-4
performance gain 8.202950000000001E-4
recognition models 8.20231E-4
generation task 8.17082E-4
poor performance 8.13669E-4
numerical attributes 8.128289999999999E-4
incremental algorithm 8.11477E-4
locative algorithm 8.099299999999999E-4
performance degradation 8.083690000000001E-4
competitive performance 8.083690000000001E-4
ing amazon 8.07292E-4
