Simon Lacoste-Julien

Simon Lacoste-Julien is a Full Professor at the Université de Montréal's Department of Computer Science and Operations Research, part of the Faculty of Arts and Sciences. He is affiliated with the Centre de recherches mathématiques (CRM) and Mila – Quebec AI Institute. His research focuses on machine learning, deep learning, optimization, and theoretical foundations of AI. Notable contributions include advancements in constrained optimization algorithms, game-theoretic learning, and sparsity-aware models.

• Education: Doctorate in Machine Learning (PhD), specialized in optimization and graphical models.

Research Interests: His work bridges theory and practice, exploring topics such as:
- Optimization for Machine Learning: Developing efficient algorithms for large-scale problems.
- Game-Theoretic Approaches: Applying game theory to prediction and mechanism design.
- Disentangled Representations: Learning structured latent variable models for better interpretability.
- Adaptive Gradient Methods: Improving convergence and stability of training processes.

Grants & Projects: Lead researcher on multiple grants including:
- Model-based Machine Learning with Guarantees (2025-2031, CRSNG)
- Robust and Efficient Structured Prediction (2017-2025)
- Projects on few-shot learning and algorithmic fairness funded by MITACS and IVADO.

Labs & Teams: Active member of Mila, leading research in optimization and theoretical aspects of deep learning. Collaborates widely with industry (e.g., Samsung Electronics) and contributes to open-source tools like the Cooper library for constrained optimization.