answer type 0.0046591
answer ranking 0.004367106
answer candidate 0.004357232
correct answer 0.004298266
answer types 0.004260469
multiple answer 0.004242402
answer extraction 0.004239538
lexical answer 0.004224458
long answer 0.004153418
possible answer 0.004151711
answer validation 0.004130567
answer candidates 0.004106185
based answer 0.00410207
answer outputs 0.004075661
ent answer 0.004071871
answer coverage 0.004053349
answer resolution 0.00404803
answer generation 0.004046
answer typing 0.004044734
unique answer 0.004033536
combined answer 0.004033339
answer candi 0.004024535
wrong answer 0.0040190140000000004
answer candidateak 0.004016664
answer typ 0.004016664
heterogenous answer 0.004016664
answer levels 0.004016664
answer 0.00372996
question types 0.002370639
factoid question 0.00230421
question focus 0.00224389
such questions 0.002207881
question answering 0.002197546
web questions 0.001962851
ranking features 0.0019524059999999998
factoid questions 0.00187802
question 0.00184013
metric questions 0.00177936
ranking model 0.0017634859999999999
scale questions 0.001713139
correlation features 0.0016825249999999998
ranking feature 0.0016677559999999998
mbr model 0.001629571
type candidate 0.001556412
such information 0.001515361
system mbr 0.001506461
component model 0.001465043
data set 0.001447591
mixture model 0.001438692
correct answers 0.001420654
model distributions 0.0014168919999999999
questions 0.00141394
single system 0.001405434
agreement feature 0.001404135
correlation feature 0.0013978749999999998
different score 0.001387678
corresponding feature 0.001386665
indicator feature 0.0013722959999999998
feature weights 0.0013676019999999999
evaluation data 0.001360237
different types 0.001353811
system indicator 0.0013449159999999998
feature hsys 0.0013201929999999999
feature hrank 0.0013201929999999999
feature hcons 0.0013201929999999999
ensemble feature 0.0013201929999999999
features 0.00131526
search space 0.001292737
ith system 0.0012918539999999998
swering system 0.001289862
system qai 0.001289862
training process 0.001262459
second approach 0.001244709
first approach 0.001239673
second data 0.0012366880000000001
first data 0.001231652
different domains 0.001208967
data sets 0.0012007279999999999
mbrar approach 0.0011791129999999999
model 0.00112634
ranking components 0.001126015
specific ranking 0.001106696
ranking results 0.001091612
language processing 0.001089442
ranking models 0.001083825
high ranking 0.001080376
combined search 0.001071003
online search 0.001069337
search engines 0.001063758
commercial search 0.00105873
search spaceh 0.00105873
search spacehi 0.00105873
passage retrieval 0.001051132
ranking scores 0.001050192
work mbr 0.0010376
feature 0.00103061
ranking accuracy 0.0010304210000000001
same ner 0.001028185
leverage information 0.00101033
other candidates 0.001007111
