word alignment 0.00518262
word translation 0.0049066000000000005
translation model 0.00417218
word alignments 0.00394367
alignment models 0.00367003
translation models 0.0033940100000000003
english word 0.0033774829999999997
hmm word 0.003274587
english alignment 0.0031241029999999996
word translations 0.0031011479999999998
word type 0.003052621
distinct word 0.003034976
word trans 0.003023717
french word 0.003013229
word classes 0.002999711
automatic word 0.0029945469999999997
glish word 0.002987967
alignment error 0.0028488669999999997
english translation 0.0028480830000000004
alignment direction 0.002804557
translation performance 0.002791782
alignment precision 0.0027877419999999997
alignment quality 0.002782545
translation table 0.002776886
language model 0.002757249
alignment accuracy 0.002748241
ibm model 0.002711962
translation rules 0.002649105
translation system 0.00264182
machine translation 0.002618135
translation experiments 0.00260054
statistical translation 0.0025644490000000003
hmm model 0.002540167
translation systems 0.0025373310000000003
good translation 0.002515699
translation prob 0.002512114
translation workshop 0.0025121120000000004
translation quality 0.0025065250000000003
translation sys 0.0024727710000000003
translation accuracy 0.002472221
alignment 0.00246462
translation pipelines 0.0024581150000000003
tion model 0.002447919
translation 0.0021886
model 0.00198358
language models 0.001979079
source words 0.00194056
ibm models 0.0019337920000000001
same data 0.001891697
other models 0.001887722
parallel data 0.001868093
baseline alignments 0.001715088
standard alignments 0.0016744400000000001
many models 0.0016352200000000002
task data 0.001633628
available data 0.00158711
english sentence 0.001585557
training method 0.001555038
development data 0.001553649
newsgroup data 0.001527961
extra alignments 0.0015194750000000002
gold alignments 0.0015162810000000002
symmetrized alignments 0.001507684
distortion models 0.0014793220000000001
rare words 0.0014704899999999999
other language 0.001455981
training set 0.001431385
different language 0.001414478
unrelated words 0.001414404
lion words 0.00141086
uncommon words 0.001409732
ibm bleu 0.001320608
standard training 0.001283806
likelihood training 0.001240312
alignments 0.00122567
french sentence 0.001221303
baseline ibm 0.0012178
models 0.00120541
discriminative training 0.001193487
language pairs 0.001176805
training procedure 0.0011614400000000001
test set 0.001159653
language pair 0.001158808
source extension 0.001142883
words 0.00114012
search algorithm 0.0011333939999999998
efficient training 0.001119734
heuristic parameters 0.001074625
ibm mod 0.0010706259999999999
ferent language 0.001054452
standard set 0.001045119
ond language 0.001043006
other con 0.0010386990000000001
small set 0.001019786
objective function 0.001018068
english string 0.001013442
simple extension 0.0010101189999999999
same number 0.001003509
scalable method 9.93846E-4
default set 9.86604E-4
