Relational data represent relationships between entities anywhere on the web (e.g. online social networks) or in the physical world (e.g. structure of the protein).
Graph neural networks are intimately related to partial differential equations governing information diffusion on graphs. Thinking of GNNs as PDEs leads to a new broad class of graph ML methods.
TL;DR: Have you even wondered what is so special about convolution? In this post, I derive the convolution from first principles and show that it naturally emerges from translational symmetry. During…
F. Wu, A. Souza, T. Zhang, C. Fifty, T. Yu, and K. Weinberger. Proceedings of the 36th International Conference on Machine Learning, volume 97 of Proceedings of Machine Learning Research, page 6861--6871. PMLR, (09--15 Jun 2019)
P. Heim, J. Ziegler, and S. Lohmann. Proceedings of the International Workshop on Interacting with Multimedia Content in the Social Semantic Web (IMC-SSW 2008), volume 417 of CEUR Workshop Proceedings, page 49--58. Aachen, (2008)
R. Dindokar, N. Choudhury, and Y. Simmhan. IEEE International Workshop on Parallel and Distributed Computing for Large Scale Machine Learning and Big Data Analytics (ParLearning), Co-located with IPDPS, page 1185--1190. (2015)Short paper.