Deadline: 30 jun 2021
CALL FOR ONE-YEAR POSTDOC ON “SPARSE CODING FOR BIO-INSPIRED VISION MODELS”
Location : L2S, CentraleSupélec, Gif-sur-Yvette, France
Applications are invited for a one-year Postdoc position, starting in September 2021 or as soon as possible thereafter, to work under the joint supervision of Dario Prandi (CNRS, Université Paris-Saclay), Luca Calatroni (CNRS, Université Côte d’Azur), and Laurent Perrinet (CNRS, Institut de Neurosciences de la Timone).
The position is financed by the ANR Rubin-Vase (https:/anr.frProjet-ANR-20-CE48-0003), and the research focus will be on linking the bio-inspired models studied by D. Prandi and L. Calatroni (see e.g. 1) to the sparsity promoting neural architectures investigated by L. Perrinet (see, e.g., 2).
The successful applicant will have a strong applied mathematical background, ideally in the fields of sparse optimization andor dynamical systems, and a good familiarity with neural architectures and the associated implementation issues. Strong skills in coding and programming using standard optimizationdeep learning languages (e.g., Python, PyTorch, Julia, MATLAB…) are required.
Although the position is officially based at CentraleSupélec (Université Paris-Saclay), the successful applicant is expected to spend fairly long periods at the INT in Marseille and I3S in Sophia-Antipolis, to work with the co-supervisors.
Candidates are asked to write directly to Dario Prandi (firstname.lastname@example.org), attaching a CV, a short description of current research interests, and names and contact addresses of at least two reference contacts (no formal letters of recommendation are required).
Applications are reviewed continuously, until positions are filled, with a first evaluation round starting on the 14th of June, 2021. A National Security clearance is needed, and it can require approximately 2 months.
Keywords : Sparse representation, neural computation, efficient coding
1 Bertalmío, M., Calatroni, L., Franceschi, V. et al. Cortical-Inspired Wilson–Cowan-Type Equations for Orientation-Dependent Contrast Perception Modelling. J Math Imaging Vis 63, 263–281 (2021). https:/link.springer.comarticle10.1007s10851-020-00960-x
2 Boutin V, Franciosini A, Chavane F, Ruffier F, Perrinet L (2021) Sparse deep predictive coding captures contour integration capabilities of the early visual system. PLoS Comput Biol 17(1): e1008629. https:/doi.org10.1371journal.pcbi.1008629