tech,01P051067,bq 
statistical MT architectures
</term>
.
<term>

Syntaxbased statistical machine translation ( MT )

</term>
aims at applying
<term>
statistical

#9408
Error analysis suggests several key factors behind this surprising finding, including inherent limitations of current statistical MT architectures.Syntaxbased statistical machine translation ( MT ) aims at applying statistical models to structured data. 
tech,101P051067,bq 
translation ( MT )
</term>
aims at applying
<term>

statistical models

</term>
to
<term>
structured data
</term>
. In

#9418
Syntaxbased statistical machine translation (MT) aims at applyingstatistical models to structured data. 
other,131P051067,bq 
applying
<term>
statistical models
</term>
to
<term>

structured data

</term>
. In this paper , we present a
<term>

#9421
Syntaxbased statistical machine translation (MT) aims at applying statistical models tostructured data. 
tech,72P051067,bq 
</term>
. In this paper , we present a
<term>

syntaxbased statistical machine translation system

</term>
based on a
<term>
probabilistic synchronous

#9431
In this paper, we present asyntaxbased statistical machine translation system based on a probabilistic synchronous dependency insertion grammar. 
other,152P051067,bq 
translation system
</term>
based on a
<term>

probabilistic synchronous dependency insertion grammar

</term>
.
<term>
Synchronous dependency insertion

#9439
In this paper, we present a syntaxbased statistical machine translation system based on aprobabilistic synchronous dependency insertion grammar. 
other,03P051067,bq 
dependency insertion grammar
</term>
.
<term>

Synchronous dependency insertion grammars

</term>
are a version of
<term>
synchronous

#9445
In this paper, we present a syntaxbased statistical machine translation system based on a probabilistic synchronous dependency insertion grammar.Synchronous dependency insertion grammars are a version of synchronous grammars defined on dependency trees. 
other,83P051067,bq 
insertion grammars
</term>
are a version of
<term>

synchronous grammars

</term>
defined on
<term>
dependency trees
</term>

#9453
Synchronous dependency insertion grammars are a version ofsynchronous grammars defined on dependency trees. 
other,123P051067,bq 
synchronous grammars
</term>
defined on
<term>

dependency trees

</term>
. We first introduce our
<term>
approach

#9457
Synchronous dependency insertion grammars are a version of synchronous grammars defined ondependency trees. 
tech,44P051067,bq 
trees
</term>
. We first introduce our
<term>

approach

</term>
to inducing such a
<term>
grammar
</term>

#9464
We first introduce ourapproach to inducing such a grammar from parallel corpora. 
other,94P051067,bq 
<term>
approach
</term>
to inducing such a
<term>

grammar

</term>
from
<term>
parallel corpora
</term>

#9469
We first introduce our approach to inducing such agrammar from parallel corpora. 
lr,114P051067,bq 
inducing such a
<term>
grammar
</term>
from
<term>

parallel corpora

</term>
. Second , we describe the
<term>
graphical

#9471
We first introduce our approach to inducing such a grammar fromparallel corpora. 
tech,55P051067,bq 
corpora
</term>
. Second , we describe the
<term>

graphical model

</term>
for the
<term>
machine translation

#9479
Second, we describe thegraphical model for the machine translation task, which can also be viewed as a stochastic treetotree transducer. 
other,95P051067,bq 
<term>
graphical model
</term>
for the
<term>

machine translation task

</term>
, which can also be viewed as a
<term>

#9483
Second, we describe the graphical model for themachine translation task, which can also be viewed as a stochastic treetotree transducer. 
tech,205P051067,bq 
</term>
, which can also be viewed as a
<term>

stochastic treetotree transducer

</term>
. We introduce a
<term>
polynomial

#9494
Second, we describe the graphical model for the machine translation task, which can also be viewed as astochastic treetotree transducer. 
tech,36P051067,bq 
transducer
</term>
. We introduce a
<term>

polynomial time decoding algorithm

</term>
for the
<term>
model
</term>
. We evaluate

#9501
We introduce apolynomial time decoding algorithm for the model. 
tech,96P051067,bq 
time decoding algorithm
</term>
for the
<term>

model

</term>
. We evaluate the
<term>
outputs
</term>

#9507
We introduce a polynomial time decoding algorithm for themodel. 
tech,37P051067,bq 
<term>
model
</term>
. We evaluate the
<term>

outputs

</term>
of our
<term>
MT system
</term>
using

#9512
We evaluate theoutputs of our MT system using the NIST and Bleu automatic MT evaluation software. 
tech,67P051067,bq 
evaluate the
<term>
outputs
</term>
of our
<term>

MT system

</term>
using the
<term>
NIST and Bleu automatic

#9515
We evaluate the outputs of ourMT system using the NIST and Bleu automatic MT evaluation software. 
measure(ment),107P051067,bq 
our
<term>
MT system
</term>
using the
<term>

NIST and Bleu automatic MT evaluation software

</term>
. The result shows that our
<term>

#9519
We evaluate the outputs of our MT system using theNIST and Bleu automatic MT evaluation software. 
tech,58P051067,bq 
</term>
. The result shows that our
<term>

system

</term>
outperforms the
<term>
baseline system

#9532
The result shows that oursystem outperforms the baseline system based on the IBM models in both translation speed and quality. 