Article Source

• The Power of Graph Learning

The Power of Graph Learning

• Floris Geerts (University of Antwerp)
• Richard M. Karp Distinguished Lecture
• https://simons.berkeley.edu/events/power-graph-learning-richard-m-karp-distinguished-lecture

Abstract

The field of graph learning is currently experiencing a surge in popularity due to its applications in areas such as drug discovery, social networks, and recommendation systems. Unlike traditional learning, graph learning requires predictions to remain consistent under the application of graph isomorphisms. This is achieved through the use of graph invariant operations, such as neighborhood aggregation and pointwise (nonlinear) function applications, in the development of graph learning methods. In this presentation, Floris Geerts will delve into how the compositional and invariant nature of graph learning methods makes them well suited for a theoretical analysis of their expressive power. He will show that, by drawing connections between graph learning and database theory, we can gain valuable insights into the expressive power potential of graph learning methods.

Bio

Floris Geerts is a professor at the University of Antwerp in Belgium. He has held research positions at the University of Edinburgh and University of Helsinki. His research interests include database theory, data quality, provenance, and, more recently, graph learning. He has been recognized for his work in data mining and data quality, received the ACM PODS Alberto O. Mendelzon Test-of-Time Award for his work on XML, and was a recipient of the prestigious ICLR Outstanding Paper Award for his work on graph neural networks.

The Richard M. Karp Distinguished Lectures were created in Fall 2019 to celebrate the role of Simons Institute Founding Director Dick Karp in establishing the field of theoretical computer science, formulating its central problems, and contributing stunning results in the areas of computational complexity and algorithms. Formerly known as the Simons Institute Open Lectures, the series features visionary leaders in the field of theoretical computer science, and is geared toward a broad scientific audience.