translation model 0.00421925
target word 0.003831264
alignment model 0.00352492
language translation 0.00335045
word context 0.0033221379999999997
word error 0.003255006
word penalty 0.003173723
translation approach 0.003129502
normal word 0.003124202
word order 0.003120037
perfect word 0.00311053
word tri 0.003084323
translation models 0.0030532
language model 0.0030107800000000002
speech translation 0.0028471010000000003
translation task 0.002838235
translation system 0.0028188500000000003
same translation 0.002806115
machine translation 0.00269164
reference translation 0.002688316
direct translation 0.002669738
whole translation 0.0026206370000000003
translation research 0.002618433
translation hypothesis 0.002608539
translation systems 0.002602216
translation quality 0.0025967300000000002
entropy translation 0.002585121
translation mod 0.002558487
symmetric translation 0.00254997
tistical translation 0.002532201
other model 0.002508845
lexicon model 0.002456178
model parameter 0.002399058
alignment models 0.00235887
model parameters 0.002352501
tion model 0.002322171
template model 0.0023180510000000002
translation 0.00227946
lation model 0.002256331
model scaling 0.002247351
tropy model 0.002194003
model pθˆ 0.002191048
acoustic model 0.002191048
language sentence 0.00215592
statistical alignment 0.002120098
target sentence 0.0020836839999999997
source language 0.00207793
target language 0.002069744
additional features 0.001971821
alignment template 0.001963391
model 0.00193979
alignment templates 0.001884045
alignment mapping 0.001862963
lexical features 0.001847166
language models 0.00184473
alignment tem 0.0018422599999999999
training corpus 0.001807996
additional language 0.001757321
baseline features 0.001714542
test sentence 0.0016804720000000001
sentence pair 0.001649675
different probability 0.001632585
grammatical features 0.001629861
feature function 0.0016219989999999998
new features 0.001594769
alignment 0.00158513
semantic features 0.001583393
source position 0.001561173
language sentences 0.00156091
target position 0.001552987
complex features 0.00154855
features logpr 0.0015454050000000001
discriminative training 0.00154512
pragmatic features 0.001538132
features yields 0.001538132
language phrases 0.001531337
parallel training 0.0015286990000000001
following feature 0.0015157859999999999
sentence error 0.001507426
reference sentence 0.001493786
target sentences 0.001488674
possible target 0.001474136
training criterion 0.001470578
similar sentence 0.001468866
training methods 0.001436895
output sentence 0.001428481
dictionary feature 0.0014193349999999999
penalty feature 0.0014085629999999998
feature functions 0.0013947019999999998
entropy training 0.001392311
subjective sentence 0.0013836270000000001
natural language 0.0013834189999999999
sentence length 0.001380649
new feature 0.0013766289999999999
gram language 0.001369917
search algorithm 0.0013653229999999999
length feature 0.001363069
trigram language 0.001351563
training cor 0.0013501189999999999
language mod 0.001350017
