Research

Jimmy Lei Ba is a third-year PhD student in the University of Toronto Machine Learning group. He obtained BASc, MASc degrees from the University of Toronto and has previously worked with Ruslan Salakhutdinov and Brendan Frey. His primary research interests are in the areas of deep neural networks. In particular, he has been developing attention-based machine learning models in the applications of computer vision. Also, He is broadly interested in questions related to computational cognitive science, numerical optimization and Bayesian statistics.

Research Summary

Jimmy Ba designs learning algorithms to train deep neural networks. His aim is to discover learning procedures beyond the standard back-propagation algorithm that can scale up neural networks to reason about complex high dimensional visual inputs. His current research on attention-based neural networks has shown that we can substantially improve the computational efficiency of deep neural networks as well as obtain a meaningful insight into the network’s behavior. Advances in image understanding with visual attention models, especially the caption/image generation system, allow further development of more powerful machine intelligence tools. His other interests to neural network research include model compression and optimization algorithms.