Offre en lien avec l’Action/le Réseau : – — –/– — –

Laboratoire/Entreprise : GREYC (Caen) or LIFAT (Tours)
Durée : 6 to 9 months
Contact : nicolas.ragot@univ-tours.fr
Date limite de publication : 2025-04-22

Contexte :
French ANR CoDeGNN project (https://www.normastic.fr/projet-anr-codegnn/)

Sujet :
The subject of the postdoc is relatively open, depending on the experience of the candidate and the interests of the research team in the field. We currently expect to work on the following topics:
‒ Pooling operations adapted to dynamic graphs, as a continuation of the work realized for static
graphs during the CoDeGNN project
‒ Dynamic graph representations and autoencoders in the context of:
* anomaly detection on time series signals, based for example on benchmark datasets like MSL, SMAP, SMD…
* analysis and synthesis of 2-team sport games, with a 1st focus on the generation of game sub-sequences from the previous sub-sequences.
‒ Dynamic graph representations for spatiotemporal time series prediction on environmental data (air pollutants, groundwater level).
More details are given on request.

A candidate with its own subject on GNNs for dynamic graphs can also apply.

Profil du candidat :
‒ Ph.D. in Computer Science, Data Science, Electrical Engineering.

Formation et compétences requises :
‒ Required experience: GNN and dynamic graphs, computer vision and pattern recognition.
‒ Python, Pytorch or Tensorflow programming.

Adresse d’emploi :
The postdoc will take place in one of the following 2 laboratories:
‒ LIFAT, Tours, France
‒ GREYC, UMR CNRS – ENSICAEN – UNICAEN, Caen, France

Document attaché : 202503022203_Postdoc_position_GNN_Dynamic_Graphs.pdf