2006-07-24: Post-doc job offer at for "Interactive and Visual Graph Mining" at INRIA / Bordeaux, France

From InfoVis:Wiki
Jump to navigation Jump to search

INRIA, the national institute for research in computer science and control, operating under the dual authority of the Ministry of Research and the Ministry of Industry, is dedicated to fundamental and applied research in information and communication science and technology (ICST).

INRIA is offering a 1 year post-doc position in Bordeaux, France in the area of "Interactive and Visual Graph Mining". You can find more information on INRIA's web site www.inria.fr (follow the link Work & Training or go directly to the URL www-futurs.inria.fr/emplois/fiches_post_doc/Fiche_Melancon_Ang.html ).

Post-doc job offer art INRIA / Bordeaux, France[edit]

Interactive and Visual Graph Mining. Applications to strategic and competitive watch, and to social network analysis.

By nature, strategic watch as well as social network analysis (in a wider sense) mix data from various sources. The goal is then to be able to discern relevant signals in massive information in order to anticipate tendencies or events that can be critical to organizations. In these contexts, interactive and visual graph mining is a fruitful approach requiring the development of techniques astutely combining dynamic graphical representations with interactive mining.

The project will evolve in collaboration with the members of the “Visualization of Massive Data” research theme of the MABioVis team at LaBRI (Université Bordeaux I). Its focus will primarily be the study of social networks (close to the on-going ANR project SPANGEO – SPAtial Network in GEOgraphy) and strategic watch (ANR RNTL grant proposal). Other application domains include biology and bio-informatics, where various data sources are combined (proteomics, clinical data, bibliographical data) in order to help biologists gain perspective and iteratively refine their research strategies.

Mining large and complex graphs to identify sub-communities requires developing ideas emerging from graphical statistics, combinatorial analysis (graphs), classification and clustering, often with the necessity to integrate specific domain knowledge. Interacting on graphical representations (and on the underlying data), and visually detecting structural patterns provide insights that sometimes cannot be captured by classical statistical analysis. When dealing with dynamic data, visually identified components or specific data elements can be put under surveillance to follow their evolution and help diagnose or decide of actions to take.

The Bordeaux team develops a software framework (Tulip www.tulip-software.org) for the visualization and manipulation of graphs, capable of dealing with large graphs, and targeted at computing and navigating graph hierarchies.

Required skills[edit]

Skills are required in visual mining (graphical representations and interaction), but also in combinatorics and/or statistics (with an emphasis on clustering/classification). The candidate must hold a Ph.D. in Computer Science and have clearly shown her/his abilities to conduct research at an international level.

Techniques developed within the team are naturally experimented with Tulip before going into future releases. We seek persons having programming experience, to compute and/or experiment with graphs, and also to acquire and/or extract data.


For more information, contact:

GUY MELANÇON (Guy.Melancon(at)lirmm.fr) Tél. +33 467 41 85 43

MAYLIS DELEST (Maylis.Delest(at)labri.fr) Tél. +33 540 00 60 82

LaBRI UMR 5800 (Université Bordeaux I) www.labri.fr



By Guy Melancon via Infovis mailing list