learning algorithm 0.002220554
inference algorithm 0.00175859
viterbi algorithm 0.001667839
algorithm updates 0.0016431290000000001
algorithm samples 0.0016200610000000001
online algorithm 0.0016169720000000002
crf model 0.0016146989999999998
specific algorithm 0.001608346
batch algorithm 0.001605501
tron algorithm 0.0016009750000000001
algorithm samplesk 0.00158643
algorithm benefits 0.0015803430000000001
feature set 0.0014569589999999999
learning approach 0.001454477
random method 0.0014351540000000001
training words 0.00143439
current model 0.001431802
other words 0.001405472
trained model 0.001397552
first method 0.00139747
crf method 0.001389057
feature vector 0.0013677149999999998
words results 0.0013604939999999999
algorithm 0.00135606
learning algorithms 0.001336093
same training 0.0012580970000000001
training set 0.00125285
second method 0.001244327
active learning 0.0012198250000000001
delta method 0.001204752
new data 0.001204264
learning rule 0.001199654
machine learning 0.0011947470000000001
output models 0.0011924
method minconfmargin 0.001189882
vector distribution 0.0011838320000000001
features vector 0.001183112
single words 0.001165429
additional words 0.001163142
third method 0.001163099
gaussian models 0.001160966
ing set 0.0011596269999999999
probabilistic models 0.001157332
online method 0.00115712
other methods 0.00115496
sequence learning 0.0011505220000000002
labeled words 0.001137369
method kbv 0.001135227
unlabeled data 0.00113144
method wkbv 0.001124866
model 0.00112185
fourth method 0.001121579
sixth method 0.001119247
fifth method 0.001119247
single word 0.0011159359999999999
confidence methods 0.001107856
tive learning 0.001105807
similar approach 0.001101803
standard methods 0.001099823
training algorithms 0.001096199
feature mapping 0.0010924939999999998
such information 0.001090987
learning community 0.001086905
beled words 0.001084507
joint feature 0.001084217
crf performance 0.001080957
ing confidence 0.001079955
erroneous words 0.001076848
individual words 0.001072405
large set 0.0010720460000000001
same number 0.00107175
test data 0.001068812
neighbor words 0.00106774
feature generation 0.001066582
first methods 0.00106054
probability distribution 0.001058432
binary set 0.00105332
crf methods 0.001052127
correct words 0.001051395
word level 0.001045267
labeled data 0.00104427
unlabeled training 0.001039349
data sets 0.001037332
word index 0.001036723
neous words 0.001034471
mative words 0.001031729
pected words 0.001031729
data points 0.0010293400000000001
erroneous word 0.001027355
distribution main 0.00102696
classification problem 0.001017281
word con 0.0010146159999999999
specific word 0.001012583
confidence information 0.001012355
different number 0.001012324
confidence value 0.001011295
confidence values 9.95559E-4
target word 9.925889999999999E-4
confidence parameter 9.88571E-4
single set 9.83889E-4
