Siamak Ravanbakhsh

Siamak Ravanbakhsh is an Assistant Professor at McGill University’s School of Computer Science and a Canada CIFAR AI Chair at Mila. His research focuses on machine learning, particularly representation learning with geometric and probabilistic principles. He holds affiliations at both McGill and Mila, with offices in McConnell Engineering Building (MC 325) and Mila (I.9).

Education: Ph.D. and M.Sc. from the University of Alberta (2015, 2009), B.Sc. from Sharif University. Previously held positions at UBC and a postdoc at Carnegie Mellon University’s Machine Learning Department.

Research interests include geometric deep learning, probabilistic models, symmetry exploitation in AI, and applications in science. Recent work emphasizes equivariant neural networks and diffusion models for tasks like anomaly detection and goal-reaching in robotics.

Publications span 20+ top venues (ICML, NeurIPS, ICLR), with notable contributions to equivariant architectures and physics-informed networks. His work has been recognized with awards for his dissertation and thesis.

Advises over 20 students (Ph.D., M.Sc.) and collaborates on projects like EqR for reinforcement learning and crystal structure analysis. Active in teaching (COMP 588, 451, 551) and has contributed to grants in AI and astrophysics.