Full Professor, Faculty of Science and Engineering, Université Laval
Mario Marchand, professor in the Department of Computer Science and Software Engineering at Université Laval, has worked in the field of machine learning for over 30 years. His research has mainly focused on performance guarantees of learning algorithms and on algorithms to optimize these guarantees. For example, he proposed the “set covering machines” to learn by performing data compression. He also proposed learning algorithms optimizing PAC-Bayes guarantees and those based on Rademacher complexity and Vapnik and Chervonenkis dimension. He has also applied learning algorithms in the field of health such as an algorithm for the prediction of the co-receptor used by type 1 HIV, algorithms for the prediction of antibiotic resistance of bacterial genomes, and algorithms for the prediction of peptide bioactivity. He is currently working on the design of fair learning algorithms in the insurance industry and on the design of algorithms that generate interpretable predictive models.