evaluation metrics 0.00316437
semantic metrics 0.00312773
evaluation metric 0.0029804000000000002
translation system 0.00297883
human evaluation 0.00293983
machine translation 0.002497213
translation quality 0.00237177
translation adequacy 0.002349371
training data 0.00227845
translation accuracy 0.002272812
automatic evaluation 0.002258496
translation decisions 0.002225668
translation sys 0.002165681
semantic role 0.002164379
translation utility 0.002162893
ation metrics 0.002144315
human adequacy 0.002101001
language model 0.002096778
semantic information 0.00208677
semantic frame 0.001997927
smt system 0.001991832
lexical features 0.001979045
human evalua 0.001959713
aggregated metric 0.0019547510000000002
based evaluation 0.001951311
semantic structures 0.001942629
human judgement 0.001922615
human judg 0.001920475
human ade 0.001913208
semantic knowledge 0.001849725
semantic objective 0.0018432289999999998
subjective evaluation 0.001797799
translation 0.00178689
metrics 0.00176306
semantic frames 0.001757116
semantic informa 0.0017407739999999998
semantic roles 0.001737873
semantic pars 0.001736465
semantic objectives 0.001736465
forum data 0.0017107329999999999
data forum 0.0017107329999999999
data experiment 0.001700039
training process 0.0016792410000000002
forum bleu 0.0016679529999999998
newswire data 0.001666328
tion model 0.001660486
similarity features 0.0016573360000000001
baseline system 0.00164291
bleu nist 0.0016367839999999999
hierarchical model 0.001628096
newswire bleu 0.0016235479999999998
lation model 0.001620862
metric 0.00157909
rate training 0.001573173
system dur 0.001563032
competitive system 0.001563032
discriminative model 0.0015567200000000002
model paradigm 0.00155538
adequate translations 0.0015509809999999999
smt systems 0.0015430980000000001
bleu andworse 0.001524669
tic features 0.001512347
valid translations 0.001476331
quate translations 0.001476331
meaningful translations 0.001476331
preferable translations 0.001476331
objective function 0.001456179
input sentence 0.001421042
tive function 0.001402513
evaluation 0.00140131
function attempts 0.001368525
jective function 0.001367655
informal language 0.001366458
tuning smt 0.001331276
baseline smt 0.001250862
meant score 0.001249768
lexical choices 0.001241728
smt pipeline 0.0012366669999999999
automatic seman 0.001229257
mantic role 0.00122441
role fillers 0.0012191720000000001
role labels 0.001217844
bias smt 0.001200679
test sets 0.001198752
system 0.00119194
model 0.00117938
large scale 0.001178303
velopment set 0.0011758530000000001
smt systemswas 0.0011726879999999999
role confu 0.00117168
forum text 0.001153044
recent surge 0.0011519030000000001
features 0.00111891
newswire text 0.001108639
adequacy judgment 0.001104042
translations 0.001102
training 0.00108566
event structure 0.001084967
error rate 0.00108269
tical machine 0.001081787
