Le contenu est en anglais ci-après. En résumé, l'IMT Atlantique (anciennement Télécom Bretagne) recrute un postdoctorant pour 2 ans sur la thématique du traitement de signal sur graphe. Un profil plutôt théorique est souhaité, avec de possibles applications sur des données issues de la neuroimagerie.
Le postdoc se déroulera sur le campus de Brest, situé en bordure de mer. Une rémunération attractive est proposée, avec un cadre idéal pour avoir le temps d'approfondir un sujet (2 ans de postdoc) sans contrainte thématique précise liée au financement.
Le postdoc sera conjointement encadré par Vincent Gripon et Nicolas Farrugia. De plus amples informations sur les travaux de l'équipe encadrante sont disponibles ici : http://www.brain.bzh
IMT Atlantique has an opening for a fully-funded two-year postdoctoral fellowship no later than June 2018, to work on Graph Signal Processing and potential applications to Neuroimaging. The host team is Br.A.In. (Brain-inspired Artificial Intelligence http://www.brain.bzh), and is in the Brest Campus.
Several application domains rely on multivariate spatiotemporal signals such as transportation networks, energy network, social networks or neural networks. Under the natural assumption that the signal properties are related to the topology of the graph where they are supported, graph signal processing (GSP) offers algorithms that fruitfully leverage this relational structure, and can make inferences about these relationships when they are only partially observed. Developing GSP tools specifically intended to advance modeling and analytics of multivariate signals is a new, fertile area of investigation. In particular, the positionning of the Br.A.In. project regarding GSP is targetted towards the following issues:
(1) finding optimal graph representations and adapting machine learning approaches accordingly, (2) generalizing GSP approaches to dynamic graphs in order to fully capture the spatiotemporal complexity of multivariate signals, and finally (3) developing new approaches to enhance the understanding and interpretability of machine learning and deep learning methods using metrics derived from GSP.
We are seeking a candidate that will contribute to our research effort on GSP by proposing innovative theoretical approaches, or extensions to existing approaches, related to at least one of the aforementioned issues. Our main target application is the analysis, decoding and interpretability of Neuroimaging data (Electroencephalography EEG or functional Magnetic Resonance Imaging fMRI), but candidates with interest or background in other application domains are encouraged to apply as well.
The present postdoctoral fellowship is funded by the Sustainable Attractivity Strategy of the Brittany Region, with the goal of attracting high profile junior researchers from all over the world, offering exceptional conditions for long-term academic development. IMT Atlantique is a public technological university focusing on the training of engineers at the MSc. level and junior researchers at the PhD level. Being a technological university, IMT Atlantique benefits from a advantageous status compared to french public universities. As such, we offer very competitive salary packages, with postdoc wages corresponing to a junior assistant professor level. Successful candidates will also benefits from 49 annual paid holidays, as well as access to multiple benefits from the institutions’ union (such as campus activities or sea sports). In addition, living costs in the Brest bay area are exceptionally low, with a high quality of life, and the Brest campus of IMT Atlantique is ideally located at the sea front.
The position is open immediately and must start no later than June 2018, for a duration of 24 months. The postdoc fellowship is accompanied with overheads for travel expenses covering (a) the participation to international conferences, (b) article processing charges for open-access publications and (c) acquisition of a computation server dedicated to the postdoc project. The postdoc fellowship will be fully devoted to his research projects, and may optionally (but will not be required to) contribute to teaching activities in machine learning, computer science, signal processing or electronical engineering.
Candidates must have defended a Ph.D. related to machine learning or computer science, or signal/image processing. Candidates with a PhD in neuroimaging with solid programming skills and a strong interest for machine learning are also welcome to apply. An additional requirement pertains to the attractivity strategy related to the regional funding; in order to attract international applications, candidates must have spent a minimum 12 months outside France after June 2014.
Please send your curriculum vitæ, contacts for two references, as well as a cover letter to email@example.com
(c) GdR 720 ISIS - CNRS - 2011-2018.