Please join us for the next institute seminar this Thursday, September 12 at 11:30am in DC 1302.
We are excited to have our own Prof. Mark Crowley from the department of Electrical and Computer Engineering to kick off the second annual AI institute seminar series! Dr. Crowley will give his perspective on the AI field and discuss some intriguing projects from his group, see more details below.
Title: Artificial Intelligence on Fire!
Abstract: AI is so on fire right now. From a very well deserved Turing Award for the godfathers of the Deep Learning revolution to endless new results and applications coming out every day, the past year has been an eventful one for AI. The general public now regularly talks about advances in Artificial Intelligence and Machine Learning. In this talk, I will give a brief review of some of the most exciting news from the past year of AI from around the world and close to home. This includes the launch of our own Waterloo Artificial Intelligence Institute, so I will look back briefly at what was accomplished and talked about in the past year as well as provide a brief look forward to this year's seminars.
I will then outline a few projects from my own UWECEML lab as case-studies for how AI/ML methods can be used to harness data and knowledge to vastly improve the ability of human beings to perform complex tasks. Some questions we've posed and tried to answer are: How can we make better predictions of snow cover on Ontario roads using limited, heterogeneous data and low-quality sensors? How can we predict, model and reduce the impact of spreading forest wildfires? Can we automatically learn higher quality, more compact models of fuel combustion to enable more efficient engines and energy production?
On this last question, I will present recent results from a collaboration demonstrating for the first time how to effectively use deep neural networks to approximate a complex combustion manifold for use in simulations for designing engines. This approach provides faster predictions than current methods and uses much less memory. While our neural network structure is quite simple, we combined a number of regularization methods and oversampling techniques to compensate for some of the challenges arising in this dataset. I will explain some of these as a useful case-study for implementing deep learning in complex environments such as this.
Speaker:
Prof. Mark Crowley
Department of Electrical and Computer Engineering
University of Waterloo
Speaker Bio:
Mark Crowley is an Assistant Professor in the Department of Electrical and Computer Engineering at the University of Waterloo. He received his PhD from the University of British Columbia in 2011 and did a postdoc at Oregon State University researching Computational Sustainability problems. He is a member of the Waterloo Artificial Intelligence Institute, the Waterloo Institute for Complexity and Innovation and a Faculty Research Fellow at Element AI. His research seeks dependable and transparent ways to augment human decision making in complex domains in the presence of spatial structure, large scale streaming data, and uncertainty. His focus is on developing new algorithms within the fields of Reinforcement Learning, Deep Learning, Manifold Learning, and Ensemble Methods. Dr. Crowley often works in collaboration with researchers and policy makers in diverse fields such as sustainable forest management, ecology, autonomous driving, physical chemistry, and medical imaging.
Date and Time:
Thursday, September 12, 2019
11:30 AM - 12:30 PM
Location: DC 1302
Light refreshments will be available.
Seminar Recording - YouTube Link: https://www.youtube.com/watch?v=iwTE8lTM_OU