|(Photographer S.N. Work)|
My main research interest is reinforcement learning: representation learning, value function approximation, model learning, exploration, and all those concepts we feel are necessary to the development of generally competent agents. Recently I have also been investigating the application of universal source coding algorithms to complex, non-binary data sources.
I helped develop both the Critterbot and the Arcade Learning Environment. The Critterbot is a reinforcement learning robot built by the Reinforcement Learning and Artificial Intelligence lab at the University of Alberta. My particular contribution was the programming of the ARM microcontroller's custom operating system. The ALE is a more recent project that aims to encourage and catalyze reinforcement learning research in large, interesting, human-made domains (i.e. Atari 2600 games).
Compute/Calcul Canada is showcasing Canadian scientists whose research is or was supported by high performance computing resources, such as those provided by Westgrid and Calcul Quebec. I'm very honoured to be one of the featured scientists.
My CV is available on request.