How Powerful are Graph Neural Networks?, Keyulu Xu, Weihua Hu, Jure Leskovec, Stefanie Jegelka, 2019International Conference on Learning Representations (ICLR)DOI: 10.48550/arXiv.1810.00826 - Formally establishes the theoretical connection between the capabilities of Message Passing GNNs and the 1-WL test.
Graph Representation Learning Book (Draft), William L. Hamilton, 2020Synthesis Lectures on Artificial Intelligence and Machine Learning, Vol. 14 (Morgan and Claypool) - An online draft book providing a general overview of GNNs, including discussions on graph structural differentiation and the WL test.