Yoshua Bengio


Yoshua Bengio, the NSERC-Ubisoft research chair, has developed an algorithm that has been applied in various aspects of video games: match-making, compression and generation of animations and modeling expressions on player faces. http://www.nouvelles.umontreal.ca/udem-news/news/20110315-montreals-new-research-chair-for-artificial-intelligence.html

This research is focused on deep-learning algorithms, a rapidly growing area of research in machine learning and artificial intelligence. The Compute Canada infrastructure is used to run the learning algorithms, i.e., training models on large quantities of data. The results of this research have the potential to impact Canadians through major advances in information technology. Large companies such as Google, Microsoft, Baidu, Facebook and IBM are rapidly expanding their use of deep learning for applications such as speech recognition, object recognition, image search and online advertising. Ubisoft Montreal has started using these algorithms in their video games. There is a whole ecosystem of start-ups, many of which are in Canada (and some of which we are collaborating with), that are also in good position to take advantage of these advances. The advances in my lab would not be possible without Compute Canada’s infrastructure because we need to run hundreds of jobs in parallel over several days to test each new model.