23 May: fyi -- research post, Dublin

Index of May 2011 | Index of year: 2011 | Full index



An opportunity has arisen at short notice for a two-year postdoctoral
fellowship at the National Centre for Language Technology, School of
Computing, Dublin City University.

The project will be conducted in collaboration with Symantec Ireland
and will be focused on confidence estimation for Statistical Machine
Translation (SMT). The aims of the project will be to explore the
selection and extraction of features for estimating specific aspects
of translation quality such as document comprehensibility, and to
investigate whether system combination can benefit from these
confidence scores by highlighting strengths and weaknesses in
individual systems. The candidate will work with multiple MT systems
and with a variety of datasets including noisy user-generated content.
The project will involve some system integration, as well as
preparation and fine-tuning for evaluation (using both automatic
metrics and user studies).

The successful postdoctoral fellow will have a capacity for
independent research and will enjoy working in a collaborative
research environment. The fellow will work closely with a PhD student
on a related project investigating the use of NLP resources in SMT,
and will be required to work both in DCU and in the Symantec campus in
Dublin.

The successful candidate will have been awarded their PhD by the
start of the fellowship and will have spent no longer than 36 months
of their post-PhD career in another postdoctoral position. Their PhD
topic should be in the area of computational linguistics and natural
language processing, and a strong preference will be given to those
candidates with machine translation experience.

The position will be supported by IRCSET (http://www.ircset.ie). To
apply, please send a CV and the names of two referees to Johann
Roturier (johann_roturier@symantec.com) and Jennifer Foster
(jfoster@computing.dcu.ie) by * Monday 23rd May*.

Index of May 2011 | Index of year: 2011 | Full index