[Elsnet-list] NLP internship at Xerox Research Center Europe, Grenoble, France

Nikoulina, Vassilina vassilina.nikoulina at xrce.xerox.com
Fri Dec 5 09:23:36 CET 2014


Title: Identification of discontinuous variants of compound terms

Duration 5-6 months

Proposers:

Salah Aït-Mokhtar

Vassilina Nikoulina

Start date: January-February 2015

Description

The main theme of the internship is the identification of terms and 
concepts in domain-specific texts, with a focus on medical texts in the 
context of the EURECA project (http://eurecaproject.eu/). We have a 
dictionary-based term identification system capable of identifying 
occurrences of terms in free texts, including non-listed term variants 
(e.g. inflected or misspelled terms). The task of the internship will 
consist in contributing to the extension of types of variations that the 
term identifier can handle. In particular, the intern will work on the 
identification and normalization of discontinuous compound terms that 
are involved in specific syntactic structures (e.g. coordination), using 
distant supervision with existing domain terminologies. An example of 
discontinuous compound terms is  “abdominal distention” in the 
expression “abdominal bloating or distention”.

Requirements

The ideal candidate is a student (MSc or PhD) in computational 
linguistics, or computer science with a good background in NLP. S/he has 
a good knowledge of syntactic structures and parsing. Good programming 
skills, preferably in Java, are also required. Prior experience in NLP 
for the healthcare domain or in terminologies/ontologies is a plus.

During the internship the candidate will acquire a significant knowledge 
and practice in the use of hybrid methods for term identification, 
including distant supervision based on rich terminologies and 
ontologies. As well, s/he will work closely with researchers and 
engineers in an international research environment.

You can find more details about this offer at 
http://www.xrce.xerox.com/About-XRCE/Internships/Identification-of-discontinuous-variants-of-compound-terms



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://list.hum.uu.nl/pipermail/elsnet-list/attachments/20141205/f2aec770/attachment.html>


More information about the Elsnet-list mailing list