[Elsnet-list] Post-doctoral position on semantics at CEA LIST, France

Gaël de Chalendar Gael.de-Chalendar at cea.fr
Tue Jul 5 13:25:30 CEST 2011

Department DIASI at CEA LIST proposes a fixed-term contract (post-doctoral
level) on a research project.

Subject: Automatic Generation of Scenarios from written procedures

The use of augmented reality (AR) for maintenance and the use of robots in 
this type of work or for personal assistance need to have 3D models of the
environment and objects. One must also express the maintenance scenarios (AR) 
or action (robotics) in terms of these models. But the manual design of such 
models is extremely expensive. However, technical documentation and 
maintenance scenarios are often described in natural language, most of the 
time using a controlled language or at least very specific constructions. 
Similarly, robots action scenarios can be stated vocaly by users. It would 
therefore be desirable to (semi-)automatically convert all these elements in 
models and action plans expressed in the formal language used by the 
The proposed work in this post-doc is a collaboration between three 
laboratories of the Département Intelligence Ambiante et Systèmes Interactifs 
(DIASI) of the CEA LIST. It will use the linguistic analysis and semantic 
tools developed at LVIC to develop new methods for converting text styles and 
apply them to the AR and Robotics needs of LRI and LSI laboratories.

State of the art and technological challenges
While work on these issues exist (see references below), the locks are still
very numerous. Many technologies that are required as part of this post-doc 
are still very experimental: semantic analysis tools (Word Sense 
Disambiguation, Semantic Role Labeling,...), exploitation of knowledge about 
the world / pragmatics, links between lexical semantics and formal semantics, 
human-machine interfaces for the development of a model (corrections, 

Work Description
The post-doc will start by a quick immersion in LVIC technology and tools for
Computational Linguistics and in target technologies and tools: 3D modeling, 
RA, robotics. He will continue with a thorough exploration of the state of the 
art. Then he will propose a work plan including the construction of two 
prototypes and an evaluation plan. If time permits, and there exists an 
appropriate evaluation campaign, he will participate in this one. The post-doc 
will always stay in contact with specific members of the various laboratories 
in order to ensure that his work is consistent with the overall objectives. To 
do this, monthly coordination meetings will be held with representatives of 
the three laboratories.
The prototypes developed will allow to semi-automaticaly convert technical
documents in models and scenarios . They will use language processing tools 
and will include a graphical interface to transform text into models with as 
few manual data manipulation as possible.

Candidate profile
The ideal candidate is a PhD in computer science, specialized in natural
language processing or Semantic Web technologies (ontologies, reasoning). 
He/she will have experience of the realization of graphical interfaces, if 
possible in Qt. Knowledge of 3D modeling would facilitate the commencement of 
the post-doc.

Location, duration and remuneration
The fixed-time contract will take place in the LVIC laboratory in its new
location (from October 2011) at Nano'Innov in Palaiseau on the Saclay plateau 
in the Paris region (France). The duration is 12 months and remuneration 
around € 2200, depending on the candidate profile.

How to apply
Applications should be sent as soon as possible and will be considered until 
the position is filled.

Send resume and cover letter, and any recommendations, by e-mail to
Gael.de-Chalendar at cea.fr

Gaël de Chalendar +33 1 46 54 80 18
Centre de Fontenay-aux-Roses
Laboratoire Vision et Ingénierie des Contenus
(Vision and Content Engineering Laboratory)
Bat. 38-2 ; 18, rue du Panorama ; BP 6
92265 Fontenay aux Roses Cedex ; France

- Romaric Besançon, Gaël de Chalendar, Olivier Ferret, Faiza Gara, Olivier
Mesnard, Meriama Laïb and Nasredine Semmar. LIMA : A Multilingual Framework 
for Linguistic Analysis and Linguistic Resources Development and Evaluation.
Proceedings of the Seventh conference on International Language Resources and
Evaluation (LREC'10). May 2010.
- Eva Hanser. SceneMaker: Intelligent Multimodal Visualisation of Natural
Language Scripts. Confirmation Report. June, 2009
- Minhua Ma. Automatic Conversion of Natural Language to 3D Animation. Thesis.
University of Ulster. July 2006
- Lee M. Seversky. Real-time Automatic 3D Scene Generation from Natural 
Language Voice and Text Descriptions. State University of New York at 
- Stéphane S. Somé and Xu Cheng. 2008. An approach for supporting system-level
test scenarios generation from textual use cases. In Proceedings of the 2008 
ACM symposium on Applied computing (SAC '08). ACM, New York, NY, USA, 724-729.
- Hans Svensson & Ola Åkerberg. Development and Integration of Linguistic
Components for an Automatic Text-to-Scene Conversion System. Lund Institute of
Technology. August 2002.
- Patrick Ye. Natural Language Understanding in Controlled Virtual 
Environments. PhD thesis. University of Melbourne. Melbourne, Australia. June 

Gael de Chalendar
Centre de Fontenay-aux-Roses
Laboratoire Vision et Ingénierie des Contenus
(Vision and Content Engineering Laboratory)
Bat. 38-2 ; 18, rue du Panorama ; BP 6
92265 Fontenay aux Roses Cedex ; France
Tél.: ; Fax.:
Email : Gael.D.O.T.de-Chalendar.A at T.cea.D.O.T.fr
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.elsnet.org/pipermail/elsnet-list/attachments/20110705/8bac85d1/attachment-0001.htm 

More information about the Elsnet-list mailing list