EUVIMEDEUVIMED entdecken
Crossref

On the Explainability of Graph Convolutional Network With GCN Tangent Kernel

Xianchen Zhou, Hongxia Wang

Neural Computation · 01.01.2023 · 35 (1) · 1-26

DOI: 10.1162/neco_a_01548 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Neural Computation, online erschienen am 01.01.2023. Verfasst von Xianchen Zhou und 1 weiteren Autor:innen. Beteiligte Einrichtung: National University of Defense Technology, Changsha 410073, P.R.C. zhouxianchen13@nudt.edu.cn. Die Arbeit wurde bisher 5-mal zitiert (Crossref) und nennt 42 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Graph convolutional network (GCN) is a powerful deep model in dealing with graph data. However, the explainability of GCN remains a difficult problem since the training behaviors for graph neural networks are hard to describe. In this work, we show that for GCN with wide hidden feature dimension, the output for semisupervised problem can be described by a simple differential equation. In addition, the dynamics of the behavior of output is decided by the graph convolutional neural tangent kernel (GCNTK), which is stable when the width of hidden feature tends to be infinite. And the solution of node classification can be explained directly by the differential equation for a semisupervised problem. The experiments on some toy models speak to the consistency of the GCNTK model and GCN.

Autor:innen und Institutionen

  • National University of Defense Technology, Changsha 410073, P.R.C. zhouxianchen13@nudt.edu.cn.
  • National University of Defense Technology, Changsha 410073, P.R.C. wanghongxia@nudt.edu.cn.

Zitieren

APA

Zhou, X., & Wang, H. (2023). On the Explainability of Graph Convolutional Network With GCN Tangent Kernel. Neural Computation, 35(1), 1-26. https://doi.org/10.1162/neco_a_01548

Vancouver

Zhou X, Wang H. On the Explainability of Graph Convolutional Network With GCN Tangent Kernel. Neural Computation. 2023;35(1):1-26. doi:10.1162/neco_a_01548

BibTeX
@article{Zhou2023,
  author = {Zhou, Xianchen and Wang, Hongxia},
  title = {On the Explainability of Graph Convolutional Network With GCN Tangent Kernel},
  journal = {Neural Computation},
  year = {2023},
  volume = {35},
  number = {1},
  pages = {1--26},
  doi = {10.1162/neco_a_01548}
}
Kostenlos starten →