Levi Lelis

Associate Professor, Canada CIFAR AI Chair with Amii

I am an Associate Professor in the Department of Computing Science at the University of Alberta, Canada. Previously, I was a Professor for six years at the beautiful Universidade Federal de Viçosa, Brazil.

My research is dedicated to the development of principled algorithms to solve combinatorial search problems. We are currently focused on combinatorial search problems arising from the search for programmatic representations for sequential decision-making problems (e.g., computer programs encoding policies for solving reinforcement learning problems). I believe that the most promising path to creating agents that learn continually, efficiently, and safely is to represent the agents' knowledge programmatically. While programmatic representations offer many advantages, including modularity and reusability, they present a significant challenge: the need to search over large, non-differentiable spaces not suited for gradient descent methods. Addressing this challenge is the current focus of my work.

Opportunities

Would you like to join my group? I am looking for motivated PhD students interested in programmatic (neural or otherwise) representations. If you're interested in joining my group, contact me by email with a description of your background and why you are interested in the type of research we do.

If you are looking for recent representative papers from our group to read before contacting me, please take a look at our recent work on learning neural languages from data and on learning semantic spaces as libraries of programs.

The following talks highlight recent work we've done in our research group.