|
Research Activity:
The research will be part of a joint project of the University of Leeds and Lingenio GmbH called HyghTra which investigates possibilities of rapid construction of hybrid high quality Machine Translation systems which combine rule based and statistical methods applied to a number of European languages. A main topic of the project is to design declarative core grammars and train them using monolingual and bilingual corpora and translation information.
The candidate will contribute to working out and to statistically optimizing the grammars for Spanish, Portuguese and Dutch and the corresponding parts of the dictionaries of the translation systems. Also he or she will be expected to contribute to integrating the work done at the University of Leeds and the development at Lingenio.
IMPORTANT: researchers must not have resided or carried out their main activity in the host country >12 months within the last 3 years prior to the recruitment.
Skills and Experience:
- Broad general knowledge in Computational Linguistics, in particular
good understanding of formal grammars and machine learning methods
- Good programming skills in C/C++ or Prolog
- Skills in scripting languages (Perl/Python) are welcome
- Good understanding of rule-based and statistical machine translation
techniques
- Experience in grammar development, preferentially on the basis of
dependence grammar formalisms
- Good knowledge of Spanish, Portuguese or Dutch and ability to design and
implement corresponding dependence grammars
- Good command of written and spoken English and German
- Skills in other Germanic and Romance Languages or Slavic languages are
welcome
Education:
Masters/Diploma in Computational Linguistics, Computer Science or equivalent.
Experience:
4 years in postgraduate research or development positions.
Duration: 18 months.
Salary: in relation to the available skills and experience
Application:
Closing date for application: January 15, 2011.
General information about Lingenio can be found at
http://www.lingenio.de/English/index.htm.
Please address questions concerning the position and e-mail applications to k.eberle@lingenio.de.
The official call is available as in PDF: Lingenio_job_2010-11.pdf
|
