alignment model 0.0038152999999999998
source word 0.002915804
target word 0.002904981
ibm model 0.0027867400000000002
alignment models 0.0027039589999999997
translation probability 0.00267459
model parameters 0.00267298
first model 0.00265752
word types 0.0025966649999999997
word position 0.002567826
model space 0.002523111
optimal model 0.002520005
alignment probability 0.0025198399999999998
empty word 0.002484132
word type 0.0024773829999999997
german word 0.0024637929999999997
bulgarian word 0.0024637929999999997
hmm model 0.002457068
translation parameters 0.0024514100000000002
model sequence 0.002443361
uniform model 0.002413606
ative model 0.002370386
translation probabilities 0.0023048920000000002
machine translation 0.002256269
standard alignment 0.002203111
translation parame 0.0021800580000000003
translation distributions 0.002165597
ical translation 0.002160418
uncertain translation 0.0021493600000000003
translation probabil 0.0021493600000000003
order alignment 0.002145212
alignment error 0.002141937
statistical alignment 0.002111708
model 0.00209581
hmm alignment 0.002080748
training data 0.002039649
source words 0.001970354
target words 0.001959531
training sentence 0.001885703
translation 0.00187424
training set 0.001745088
language sentence 0.001744378
null words 0.001743695
alignment 0.00171949
ibm models 0.001675399
different training 0.001666902
source sentence 0.0016455380000000002
target sentence 0.0016347150000000001
target language 0.001541445
data size 0.001525768
coupled words 0.001520421
random models 0.001517745
parallel data 0.001517458
large training 0.001493111
small data 0.001492489
same source 0.0014800960000000002
objective function 0.001474329
sentence pair 0.001459736
parameter set 0.0014395369999999998
sentence pairs 0.001410886
order models 0.0014101909999999999
such set 0.001398882
reasonable data 0.001395323
training sets 0.001367069
linear function 0.00136294
different parameter 0.001361351
data sizes 0.001355664
lel data 0.00135504
distinct models 0.001333523
alignments 0.00133061
ent models 0.00132912
probability distribution 0.001318914
language pairs 0.0013176160000000002
test set 0.001315679
above sentence 0.001298127
probabilistic models 0.001276532
entire training 0.001267791
timal models 0.001263919
trained models 0.001260934
trastive models 0.001259377
words 0.00124364
creased training 0.001241591
mass function 0.001235666
set size 0.001231207
sentence example 0.0012266970000000001
short sentence 0.0012184420000000001
concave function 0.001199233
small set 0.001197928
function xtk 0.00119729
such experiments 0.0011957930000000001
onn sentence 0.001193616
features 0.00117612
length probability 0.001156083
corresponding target 0.001138961
first method 0.001138874
optimization problem 0.0011347380000000002
parameter values 0.001132356
single source 0.0011174890000000002
probability mass 0.001113751
source sentences 0.0011137060000000001
