From AlphaGO to ChatGPT Public Talk

Wednesday, April 12, 2023 1:30 pm - 3:00 pm EDT (GMT -04:00)

from alphago to chatgpt banner

Sponsored by the Faculty of Math Data Science Graduate Programs, please join University of Waterloo expert in artificial intelligence, Professor Pascal Poupart, for a public talk in which he will describe the key technological advances in recent years which were behind, and ultimately facilitated, these breakthroughs.

In recent years, AlphaGo beat the best human players in the challenging game of Go, large language models such as ChatGPT and GPT-4 converse in a human way while displaying an unprecedented depth of knowledge, DALL-E 2 and Stable Diffusion create realistic images and art from natural language descriptions, the auto industry is rolling out AI-based autonomous driving systems, Alpha Fold 2 predicts the structure of over 200 million proteins to accelerate scientific research and other large foundational models can predict the physical and chemical properties of compounds to accelerate the design of new materials. 

While the industry often highlights the increasing size of their models (e.g., billions to trillions of parameters), this gives a false impression that simply throwing a lot of data to large computer clusters in order to train ever larger models is the key.  However, important algorithmic advances were necessary to achieve those breakthroughs.  Professor Poupart will explain the role that residual optimization and stochastic optimization played to enable deep learning.  He will also discuss important advances in reinforcement learning, self-supervised learning and few-shot learning that improved significantly the quality of foundational models.

The success of AI has also led to new challenges in terms of explainability.  While most models in advanced AI systems are difficult to interpret, Professor Poupart will also discuss recent advances in probing and conversational agents that can provide some degree of explainability.

Coffee and tea will be provided.

Registration is short, sweet and not mandatory. However, it is strongly encouraged and appreciated so we can anticipate the number of participants and provide you with any updates and information closer to the date - thank you! 

Pascal Poupart is a Professor in the David R. Cheriton School of Computer Science at the University of Waterloo, Waterloo (Canada). He is also a Canada CIFAR AI Chair at the Vector Institute and a member of the Waterloo AI Institute. He serves on the advisory board of the AI Institute For Advances in Optimization (2022-present). He served as Research Director and Principal Research Scientist at the Waterloo Borealis AI Research Lab funded by the Royal Bank of Canada (2018-2020). He also served as scientific advisor for ProNavigator (2017-2019), ElementAI (2017-2018) and DialPad (2017-2018). He received the B.Sc. in Mathematics and Computer Science at McGill University, Montreal (Canada) in 1998, the M.Sc. in Computer Science at the University of British Columbia, Vancouver (Canada) in 2000 and the Ph.D. in Computer Science at the University of Toronto, Toronto (Canada) in 2005. His research focuses on the development of algorithms for Machine Learning with application to Natural Language Processing and Material Design. He is most well known for his contributions to the development of Reinforcement Learning algorithms. Notable projects that his research team are currently working on include Bayesian federated learning, probabilistic deep learning, data efficient reinforcement learning, conversational agents, automated document editing, sport analytics, adaptive satisfiability and CO2 conversion & capture.