Screenshot 2018 10 22 at 16.03.02

OCKHAM

Optimization, pHysical Knowledge, Algorithms and Models

Team leader: Rémi Gribonval

Keywords: Graphs, Sparsity, Large scale optimization, Structured models, Dimension reduction

 

Goals: To develop computationally efficient and mathematically founded methods and models to process high-dimensional data

Applications: Imaging science, Signal processing, Neuroscience

 

  • FRUGAL METHODS WITH ROBUST EXPRESSIVITY

Sketching

      Modern avatars of sparsity

      Learning of structured tranforms

       

        • MODEL INTEGRATION IN LEARNING ALGORITHMS

        Learning of graphs and learning on graphs

        Physics inspired neural networks

         

        • INTERPRETABILITY, EXPLAINABILITY AND PRIVACY

        Model identifiability in Machine Learning and Inverse Problems

        Privacy-preserving learning

         

        Main collaborations:

        UMPA (ENS de Lyon)
        SiSyPHe (Lab. Phys., ENS de Lyon)
        Lund University (Sweden)
        CMAP (Ecole Polytechnique)
        IRISA (Université Bretagne Sud)
        IMO (Université Paris-Saclay)
        EPFL (Switzerland)
        University of Edimburg (UK)
        UCLouvain (Belgique)
        Valeo.ai (Paris)
        IRIT (Toulouse)