Research Projects

PhD Students     Former Students       Masters      Visitors     Collaborators       Research Projects

 

Current Research Projects



Information Theory beyond Communications: Distributed Representations and Deep Learning, Marie Skłodowska-Curie Global Fellowships


  1. Overview: Deep learning is an enormously successful recent paradigm with record-breaking performance in numerous applications. The individual layers in a multilayer neural network are trained  to convert high-dimensional inputs into low-dimensional codes. This joint project is a collaborative endeavor between Prof. Pablo Piantanida (L2S - CentraleSupélec) and Prof. Yoshua Bengio (MILA - Université de Montréal - IVADO) to capitalize on powerful and fertile concepts from information theory and statistics in order to advance the state of the art in deep learning. The proposed research is expected to bridge the gap between theory and practice to facilitate a more thorough understanding and hence improved design of deep learning architectures from a theoretical (information theory and statistics) and a more practical (learning algorithm design) perspective. The results of our research will be disseminated at  flagship conferences and in prestigious journals of the field.

  2. Duration: From August 2018 to July 2020









Associated International Laboratory (LIA) of the The National Center for Scientific Research (CNRS) between France - Québec on "Information, Control and Learning - The Ingredients of Intelligent Grids"


  1. Overview: Evolving information technologies, including smart meters and smart grids, can provide companies with unprecedented capabilities for forecasting demand, shaping customer usage patterns, preventing outages and more. At the same time, these advances also generate unprecedented data volume, speed and complexity. In order to deal with and use this information to gain insight, utility companies must be capable of high-volume data management and advanced analytics designed to transform data into actionable insights. This raises fundamental questions about the interplay between information, learning and control. Besides this, the information-intensive nature of the Smart Grid introduces new security and privacy considerations as well. This LIA is a collaborative endeavor of researchers with complementary backgrounds. Its main innovation is the idea to capitalize on powerful and fertile concepts from information theory, machine learning and control theory in order to advance the state of the art in managing big data for Smart Grids. The projected is coordinated  by Prof. Pablo Piantanida.

  2. Duration: From March 2017 to December 2020

  3. Partners: Université de Montréal, University McGill, L'Institut national de la recherche scientifique (INRS), École de technologie supérieure (ETS), Université Concordia, CNRS, Université Paris Saclay, CentraleSupélec, Université Paris Sud.