Simon Lacoste-Julien

Institut national de recherche en sciences et technologies du numérique, Université de Montréal

Papers

Total Citations

H-Index

About

Simon Lacoste-Julien is a leading researcher in machine learning, with key contributions spanning optimization, probabilistic inference, and video understanding. His work on the Frank-Wolfe algorithm, particularly in "Sequential Kernel Herding," has been highly influential, demonstrating how this classic optimization technique can be repurposed for efficient particle filtering and adaptive quadrature—a method that often outperforms standard Monte Carlo approaches. This line of work has garnered significant attention, with his papers accumulating thousands of citations, reflecting their foundational impact on both theory and practice. Beyond optimization, Lacoste-Julien has advanced the field of multimodal learning, notably through "Learning from Narrated Instruction Videos," where he tackled the challenging problem of aligning visual and verbal content to enable automatic task guidance. His research bridges rigorous mathematical foundations with real-world applications, from robotics to video analysis. Recognized for his clarity and depth, Lacoste-Julien’s work continues to shape how machines learn from data and interact with complex environments, making him a pivotal figure in modern AI.