David Duvenaud

Canada Research Chair in Generative Models

Tier 2 - 2017-11-01
Renewed: 2022-04-01
University of Toronto
Natural Sciences and Engineering Research Council

Research summary

Deep learning is a subfield of machine learning that is concerned with algorithms inspired by the structure and function of the brain. Many practitioners of language processing, machine vision, robotics and predictive models are eager to apply deep-learning solutions to real-world problems. But those who work in medicine, business and the physical sciences still struggle to fit large predictive models into domains like medical histories, user interaction traces and climatological datasets.

As Canada Research Chair in Generative Models, Dr. David Duvenaud is trying to address this issue. He and his research team are building new model classes and fitting procedures that are flexible and scalable enough to "meet the data where it lives," so to speak, with little to no pre-processing so it can be used in a wider range of fields and applications. Ultimately, their work will improve the accuracy of the judgments and predictions made by both humans and machines.