Back to publications
DiGRAF: Diffeomorphic Graph-Adaptive Activation Function
Paper Details
Published: 2024/10/30
DOI: 10.48550/arXiv.2407.02013
ARXIV ID: 2407.02013v2
In this paper, the authors propose a novel activation function tailored specifically for graph data in Graph Neural Networks (GNNs), DiGRAF, leveraging Continuous Piecewise-Affine Based (CPAB) transformation. In addition to its graph-adaptivity and flexibility, DiGRAF also possesses properties that are widely recognized as desirable for activation functions, such as differentiability, boundness within the domain, and computational efficiency. The authors conduct an extensive set of experiments across diverse datasets and tasks, demonstrating a consistent and superior performance of DiGRAF compared to traditional and graph-specific activation functions, highlighting its effectiveness as an activation function for GNNs
Authors
K.S.I. Mantri
X. Wang
C.B. Schonlieb
B. Riberiro