Mario Marchand

Mario Marchand is a retired Professor in the Department of Computer Science and Software Engineering at Université Laval, Canada. His research focuses on machine learning theory, explainable AI, and computational biology. He has contributed to foundational work in PAC-Bayesian analysis, generalization bounds, and kernel methods. Marchand is affiliated with the Computer Vision and Systems Laboratory and the GRAAL research group. Though no longer supervising graduate students, his recent work addresses challenges in algorithmic stability, meta-learning, and feature attribution consensus. His publications span topics from neural network optimization to applications in bioinformatics and drug discovery.

• Research Interests: Machine Learning Theory, Explainable AI, Bioinformatics, Kernel Methods, Domain Adaptation, and Statistical Learning.
• Key Contributions: PAC-Bayesian risk bounds, decision tree analysis, and algorithms for mixture models. His work bridges theoretical guarantees and practical applications in healthcare and computational biology.
• Professional Activities: Taught courses including IFT-7002 (Foundations of Machine Learning) and contributed to international conferences. His methods are used in biomarker discovery and peptide design for drug development.

Recent trends in his articles emphasize explainable AI (XAI) and resolving feature attribution disagreements, alongside foundational studies on generalization in meta-learning and multi-source domain adaptation. His collaborative projects include predicting molecular properties and advancing neural network stability.