#10874In this paper we study a set of problems that are of considerable importance toStatistical Machine Translation ( SMT ) but which have not been addressed satisfactorily by the SMT research community.
other,30-1-E06-1004,ak
been addressed satisfactorily by the
<term>
SMT research community
</term>
. Over the last decade , a variety
#10889In this paper we study a set of problems that are of considerable importance to Statistical Machine Translation (SMT) but which have not been addressed satisfactorily by theSMT research community.
tech,8-2-E06-1004,ak
Over the last decade , a variety of
<term>
SMT algorithms
</term>
have been built and empirically tested
#10901Over the last decade, a variety ofSMT algorithms have been built and empirically tested whereas little is known about the computational complexity of some of the fundamental problems of SMT.
other,22-2-E06-1004,ak
whereas little is known about the
<term>
computational complexity
</term>
of some of the fundamental problems
#10915Over the last decade, a variety of SMT algorithms have been built and empirically tested whereas little is known about thecomputational complexity of some of the fundamental problems of SMT.
tech,31-2-E06-1004,ak
some of the fundamental problems of
<term>
SMT
</term>
. Our work aims at providing useful
#10924Over the last decade, a variety of SMT algorithms have been built and empirically tested whereas little is known about the computational complexity of some of the fundamental problems ofSMT.
other,10-3-E06-1004,ak
providing useful insights into the the
<term>
computational complexity
</term>
of those problems . We prove that
#10936Our work aims at providing useful insights into the thecomputational complexity of those problems.
model,4-4-E06-1004,ak
those problems . We prove that while
<term>
IBM Models 1-2
</term>
are conceptually and computationally
#10946We prove that whileIBM Models 1-2 are conceptually and computationally simple, computations involving the higher (and more useful) models are hard.
model,22-4-E06-1004,ak
involving the higher ( and more useful )
<term>
models
</term>
are hard . Since it is unlikely that
#10964We prove that while IBM Models 1-2 are conceptually and computationally simple, computations involving the higher (and more useful)models are hard.
other,8-5-E06-1004,ak
it is unlikely that there exists a
<term>
polynomial time solution
</term>
for any of these
<term>
hard problems
#10976Since it is unlikely that there exists apolynomial time solution for any of these hard problems (unless P = NP and P#P = P), our results highlight and justify the need for developing polynomial time approximations for these computations.
other,15-5-E06-1004,ak
time solution
</term>
for any of these
<term>
hard problems
</term>
( unless P = NP and P #P = P ) ,
#10983Since it is unlikely that there exists a polynomial time solution for any of thesehard problems (unless P = NP and P#P = P), our results highlight and justify the need for developing polynomial time approximations for these computations.
other,38-5-E06-1004,ak
and justify the need for developing
<term>
polynomial time approximations
</term>
for these computations . We also
#11006Since it is unlikely that there exists a polynomial time solution for any of these hard problems (unless P = NP and P#P = P), our results highlight and justify the need for developingpolynomial time approximations for these computations.
other,9-6-E06-1004,ak
some practical ways of dealing with
<term>
complexity
</term>
. In this paper a novel solution
#11022We also discuss some practical ways of dealing withcomplexity.