Rupam Mahmood

Rupam Mahmood develops reinforcement learning algorithms and real-time learning systems for controlling physical robots.

Continually improving robot minds

Rupam Mahmood develops reinforcement learning algorithms and real-time learning systems for controlling physical robots. His research focuses on developing general and constructive mechanisms for continually improving robot minds. Currently, he is working on two long-term programs consisting of several short-term projects: a simple and general reinforcement learning system for robot control, and core constructive mechanisms for continually learning agents. In the first program, Rupam and his team are working to develop a reinforcement learning system that can be easily deployed in many different robots for solving various tasks. The second program has his research teams developing and analyzing algorithms for learning policies and representations in a continual learning setup, in which the agent is expected to go through a series of changes in the environment and tasks. Through these two programs, Rupam seeks to develop a system that enables scientific understanding as well as large-scale industrial adoption of robotics by analyzing and addressing shortcomings of current policy and representation learning methods.

Rupam is an Assistant Professor in the Department of Computing Science at the University of Alberta and a Fellow and Canada CIFAR AI Chair at Amii. In 2017, he achieved his Ph.D. in Statistical Machine Learning at the University of Alberta under the supervision of Richard S. Sutton (Amii Fellow, Chief Scientific Advisor and a pioneer of reinforcement learning) with his thesis focusing on incremental off-policy reinforcement learning algorithms. Previously, Rupam was a Research Scientist and then Lead of AI Research at Kindred Inc., where he now acts as a Scientific Advisor. He is an Associate Editor of the IEEE/RJS International Conference on Intelligent Robots and Systems (IROS) and a Senior Program Committee Member for the International Joint Conference on Artificial Intelligence (IJCAI). Rupam has produced software that provides a computational framework and a benchmark task suite for developing and evaluating reinforcement learning methods with physical robots.