Robustness of Graph Neural Networks

Explore the robustness of Graph Neural Networks in this 56-minute webinar presented by Stephan Gunnemann from Technical University of Munich. Delve into improving robustness techniques, analyze nonadaptive attacks, and understand robustness certificates. Examine whitebox and graybox certificates, graph models, and collective reasoning. Part of the Data sciEnce on GrAphS (DEGAS) Webinar Series, this presentation offers valuable insights for researchers and practitioners in the field of graph-based machine learning and signal processing.

Syllabus

Introduction
Background
Improving Robustness
Does this simple change work
Nonadaptive attacks
Intermediate message
Robustness certificates
Whitebox certificates
Graph models
Graybox certificates
Collective reasoning
Summary