Representational Power of Graph Neural Networks - Stefanie Jegelka

Explore the representational power of Graph Neural Networks in this 28-minute lecture by Stefanie Jegelka from the Massachusetts Institute of Technology. Delivered at the Institute for Advanced Study's Workshop on Theory of Deep Learning: Where next? on October 18, 2019, delve into learning problems, message passing, and aggregation operations in graph neural networks. Examine various types of graphs, their representations, and the impact of mean aggregation through empirical examples. Investigate network structure, influence distribution, and connectivity as they relate to reasoning tasks and encoding in graph neural networks. Gain insights into the theoretical foundations and practical applications of this powerful machine learning approach for graph-structured data.