Graph neural network: Difference between revisions

Revision as of 22:34, 30 June 2021

Revision as of 22:34, 30 June 2021

Line 56:

===== GCN<ref>{{Cite journal|last=Kipf|first=Thomas N.|last2=Welling|first2=Max|date=2017-02-22|title=Semi-Supervised Classification with Graph Convolutional Networks|url=http://arxiv.org/abs/1609.02907|journal=International Conference on Learning Representations (ICLR), 2017}}</ref> =====

Given a input graph with <math>N</math> nodes and the node feature matrix <math>X</math>, GCN simplified the convolution layer to �<math>H^{(l+1)} = \sigma (\tilde{D}^{-1/2}\tilde{A}\tilde{D}^{-1/2}H^{(l)}W^{(l)})</math>, where

* <math>\tilde{A} = A + I</math>;