Skip to content

Simon Lacoste-Julien


  • Associate Fellow
  • Canada CIFAR AI Chair
  • Learning in Machines & Brains
  • Pan-Canadian AI Strategy




Simon Lacoste-Julien’s research focuses on machine learning, i.e., how to program a computer to learn from data and solve useful tasks.

His primary research goal is to develop and analyze machine learning techniques that can exploit, at large scale, the rich structure of data in interdisciplinary applications such as natural language processing, information retrieval, computer vision and computational biology. To this end, he combines tools from optimization, statistics and computer science, and particularly enjoys working at the interface between domains. Lacoste-Julien is best known for his contributions in three areas: structured prediction (classification problems where the outputs are structured objects such as sequences or graphs); large scale optimization (incremental gradient method and Frank-Wolfe optimization); and the combination of generative and discriminative methods.


  • NSERC Discovery Grant, 2017
  • Google Focused Research Award, 2016
  • Wolfson College Junior Research Fellowship, University of Cambridge, 2009–11
  • UC Berkeley College of Engineering Graduate Student Prize, 2008

Relevant Publications

  • Osokin, A., F. Bach, and S. Lacoste-Julien. "On Structured Prediction Theory with Calibrated Convex Surrogate Losses." Paper presented at NIPS conference, Long Beach, 2017.
  • Lacoste-Julien, S., and M. Jaggi. "On the Global Linear Convergence of Frank-Wolfe Optimization Variants." Paper presented at NIPS conference, Montreal, 2015.
  • Defazio, A., F. Bach, and S. Lacoste-Julien. "SAGA: A Fast Incremental Gradient Method With Support for Non-Strongly Convex Composite Objectives." Paper presented at NIPS conference, Montreal, 2014.
  • Lacoste-Julien, S. et al. "Block-Coordinate Frank-Wolfe Optimization for Structural SVMs." Paper presented at ICML conference, Atlantia, GA, 2013.
  • Lacoste-Julien, S., F. Huszár, and Z. Ghahramani. "Approximate Inference for the Loss-Calibrated Bayesian." Paper presented at AISTATS conference, Fort Lauderdale, FL, 2011.

Support Us

CIFAR is a registered charitable organization supported by the governments of Canada, Alberta and Quebec, as well as foundations, individuals, corporations and Canadian and international partner organizations.

MaRS Centre, West Tower
661 University Ave., Suite 505
Toronto, ON M5G 1M1 Canada