word alignment 0.0047576
translation model 0.00422804
word alignments 0.0034886989999999996
hmm model 0.003291173
fertility model 0.003280495
english word 0.003148518
position model 0.0031326419999999997
word position 0.0030843219999999996
jump model 0.002959313
positional model 0.002946372
word index 0.002935653
ity model 0.0029058
alignment inference 0.002905301
french word 0.0029042679999999998
term word 0.002903298
deficient model 0.002898472
seminal model 0.002898472
hmm alignment 0.002884993
word length 0.0028751049999999998
word sequences 0.0028706369999999997
translation models 0.0028658
glish word 0.002853843
word prefix 0.002850217
letter word 0.002850217
word correspondences 0.002850217
alignment error 0.0026484860000000002
model 0.00260605
alignment vector 0.00260027
prior alignment 0.002571726
alignment variable 0.002564214
alignment vectors 0.002550321
final alignment 0.002514973
alignment quality 0.0025087350000000002
likely alignment 0.002507281
hmm translation 0.002307113
alignment 0.00219987
translation distribution 0.002188105
translation task 0.002122231
machine translation 0.002077954
translation process 0.002003298
translation dis 0.001960437
translation systems 0.001960243
ibm models 0.001782161
translation 0.00162199
latter models 0.001567481
complicated models 0.001546865
training data 0.001444256
other words 0.001421238
english words 0.0014022779999999999
fertility hmm 0.001359568
posterior probability 0.001348827
viterbi algorithm 0.001342378
such distribution 0.00132186
sentence pairs 0.0013098089999999999
exact algorithm 0.001294463
sentence pair 0.001268272
log probability 0.001257625
models 0.00124381
algorithm aer 0.001241482
fertility distribution 0.00124056
decomposition algorithm 0.001233581
programming algorithm 0.001233362
training sentences 0.001217472
training com 0.001178981
long words 0.001178049
standard hmm 0.001173491
multiple words 0.001169244
exact parameter 0.001159958
french words 0.001158028
mon words 0.001151394
decomposition method 0.001149164
french sentence 0.001140945
complete probability 0.001135332
rare words 0.001128852
crucial training 0.0011246659999999999
modeling fertility 0.001120567
posterior distribution 0.0011204750000000001
objective function 0.001115422
same data 0.001109494
infrequent words 0.001109075
unaligned words 0.0011047890000000001
algorithm prac 0.001102546
dual decomposition 0.001099952
heuristic search 0.001096945
probability mass 0.001088622
probability includ 0.001088622
problem argmaxa 0.001079666
baseline hmm 0.001076466
parameter estimation 0.001076174
recent work 0.001076067
local search 0.001073294
timization problem 0.001070259
global parameter 0.001069762
inference techniques 0.001067195
linear time 0.0010644130000000002
map inference 0.001057278
inference procedures 0.0010526939999999999
such meth 0.0010516190000000002
simple set 0.001046555
gramming method 0.0010434889999999999
