transliteration system 0.00329436
transliteration model 0.00326785
training data 0.00289515
english word 0.002694025
word alignment 0.002621503
transliteration pairs 0.002543033
word pairs 0.002531553
such word 0.002471704
language pairs 0.002426653
target word 0.0024255780000000003
data probability 0.0023923449999999997
transliteration systems 0.002380433
transliteration information 0.002376054
source word 0.002369418
arabic word 0.002326573
other language 0.002307389
correct transliteration 0.002268175
unsupervised transliteration 0.002252941
supervised transliteration 0.002251244
task data 0.002247808
transliteration mining 0.002234542
transliteration pair 0.002176088
transliteration gold 0.002175882
high transliteration 0.00216831
word pair 0.002164608
exact transliteration 0.002163721
data count 0.002152382
russian word 0.002118694
transliteration infor 0.002118546
reference data 0.002113804
transliteration prob 0.002104233
word boundary 0.002092468
labelled data 0.0020870999999999997
ing data 0.00207956
noisy data 0.0020785789999999997
extracts transliteration 0.00207852
data con 0.0020778439999999997
data counts 0.0020760789999999998
word charac 0.0020750250000000003
separate word 0.00207397
word definition 0.002073522
mines transliteration 0.002070203
data probabilities 0.002068246
transliteration min 0.002066037
ation word 0.002063272
transliteration submodel 0.00206285
transliteration units 0.002061439
language pair 0.002059708
data sparseness 0.0020562889999999998
word segmenter 0.002050563
inconsistent word 0.002048713
seed data 0.0020477019999999998
data avail 0.0020361949999999998
unlabelled data 0.002030199
erence data 0.002014283
data probabil 0.002014283
unsupervised system 0.001960681
supervised system 0.001958984
mining system 0.001942282
unsupervised model 0.0019341710000000002
mining model 0.001915772
unigram model 0.0018983580000000002
training pairs 0.0018970430000000002
second model 0.0018555720000000002
same training 0.001846785
ing system 0.00183278
system con 0.001831064
current system 0.00181579
channel model 0.001795512
transliteration 0.00179331
tive model 0.001786031
novel model 0.0017740350000000002
model our 0.001773807
system gen 0.001770542
pervised system 0.00176993
eration model 0.0017581950000000002
model estimation 0.0017522590000000002
language 0.00167693
training process 0.001584336
initial training 0.0015259800000000001
system 0.00150105
target words 0.001489113
labelled training 0.00148659
model 0.00147454
such pairs 0.001439597
unlabelled training 0.001429689
arabic words 0.0013901080000000001
alignment sequence 0.001294018
same wikipedia 0.001286691
english characters 0.001280889
possible alignment 0.001269552
character models 0.001256951
ferent languages 0.001232884
same problem 0.001224691
english wil 0.001193305
edit distance 0.001192853
english noun 0.00118621
russian words 0.001182229
english counterparts 0.001180122
alignment sequences 0.0011724679999999999
