[Elsnet-list] postdoc position in NLP/machine learning, Univ. of Edinburgh

Sharon Goldwater sgwater at inf.ed.ac.uk
Tue Aug 3 16:42:20 CEST 2010


                    School of Informatics
                   University of Edinburgh
           Application closing date: 1 Sept., 2010

The School of Informatics at the University of Edinburgh invites
applications for the post of Postdoctoral Research Associate on the
project "Word Segmentation from Noisy Data with Minimal Supervision",
funded by the EPSRC and supervised by Dr. Sharon Goldwater.  The
project will design and implement novel language models for use in an
unsupervised or minimally supervised setting to simultaneously perform
word segmentation and correction of noise and variability in the
realization of words (as might occur in the output of optical
character recognition, or due to phonetic variability in speech).
Applications of this general approach are many and varied. Examples
include language acquisition (exploring how children begin to identify
words in the speech they hear) as well as OCR and spelling correction.

You will have (or be near completion) a PhD in computational
linguistics, computer science, cognitive science, mathematics, or a
related discipline.  You will have proven programming skills and a
background in natural language processing (e.g., language modelling,
finite state methods, corpus processing, evaluation).  Experience with
machine learning and probabilistic modelling and inference techniques
is essential; familiarity with issues in cognitive modelling and/or
child language acquisition is also desirable.

The starting date will be 01 Oct 2010, or as soon as possible thereafter.

Fixed Term: 3 years
Salary Scale: £29,853 - £35,646 per annum
Vacancy Reference: 3013193
Closing Date: 1st September 2010

Informal inquiries can be made by email to Sharon Goldwater
(sgwater at inf.ed.ac.uk).

For further details of this position and to submit your application,
please visit http://www.jobs.ed.ac.uk/ and enter the vacancy reference
above, or use the link below.


More information about the Elsnet-list mailing list