[Elsnet-list] Researcher position available: PhD in Natural Language Processing

Nicolas HERNANDEZ nicolas.hernandez at univ-nantes.fr
Thu Mar 21 13:08:45 CET 2013


Researcher position available: PhD in Natural Language Processing

The University of Nantes (West coast of France) offers an opening for a
3-year
PhD position at the LINA Computer Sciences Laboratory in the NLP Team (TALN).

*Subject* : "Discourse structure analysis and multilingual terminology
alignment from comparable corpora.
Toward a discourse definition of the notion of context"

The deadline for application is April 19, 2013 , but consideration of
candidates
will continue until the position is filled. It is expected to start on
October 2013.

The program will be funded by a grant from the French government.
Between 1684,93 Euros per month and 2024,70 Euros (with complementary
teaching
activity) before taxes.

*Context*
The LINA TALN group (http://www.lina.univ-nantes.fr) leads research in
several NLP domains such as term extraction,
syntactic and semantic analysis, and develops several applications (e.g.
multilingual terminology extraction, opinion mining, plagiarism detection).
The group participates in various projects funded by regional, national
and European sources.
Multilingual terminology alignment from comparable corpora is one of the
major
issue of the automatic translation problem. To tackle this issue, the
baseline
approach proposes to align terms whose contexts are considered as similar
thanks to bilingual dictionaries. This approach presents some drawbacks since
the context model is quite simple (a bag of words occurring around the
considered term) and because it requires external resources.
The current study aims at exploring a new approach for building term
contexts.
The idea is to use a more linguistically inspired approach: in particular to
use discourse analysis both for providing semantically delimited text area
around term occurrences and rhetorically dependent utterances in relation to
the utterance where a term occurs. The work will start by trying out
state-of-art discourse analysis methods, then it will go into the
definition of
a discourse context notion related to the task in depth. This research will
pursue the works accomplished in the national and European projects MeTRICC
and TTC.

*Qualifications*
The ideal candidate would have:
- (or soon receive) a master degree in Computer Science, Computational
Linguistics or related area
- a strong background in machine learning, text mining, and natural
language processing
- good programming skills (e.g., Java, Python)
- excellent English language skills (written and spoken) and ability to
learn French
- experience in open source development (appreciated)

*Application procedure*
Only applicants who have been ranked in the top tier of their Master degree
class will be considered.
Candidates interested in the position are asked to contact Nicolas Hernandez
and Emmanuel Morin (firstname.lastname at univ-nantes.fr) with the following
documents: A letter of motivation outlining your interest in the specific
project,
a curriculum vitae, academic qualifications and ranks (proof to provide as
soon
as possible), at least two recommendation letters from a senior
researcher/professor who can judge your potential as a future PhD student.

*More information on*
http://e.nicolas.hernandez.free.fr/pub/rec/13
http://www.edstim.fr/these/sujets-de-these/informatique




More information about the Elsnet-list mailing list