Kolmogorov–Arnold Graph Neural Networks

Gianluca De Carlo, Andrea Mastropietro, Aris Anagnostopoulos

Keywords: Graph Neural Networks, Kolmogorov-Arnold Networks, Interpretability

TL;DR: Adaptation of Kolmogorov-Arnold Networks to graph-structured data enhancing accuracy and interpretability.

Abstract:

Graph neural networks (GNNs) excel in learning from network-like data but often lack interpretability, making their application challenging in domains requiring transparent decision-making. We propose the Kolmogorov–Arnold Network for Graphs (KANG), a novel GNN model leveraging spline-based activation functions on edges to enhance both accuracy and interpretability. Our experiments on five benchmark datasets demonstrate that KANG outperforms state-of-the-art GNN models in node classification, link prediction, and graph classification tasks. In addition to the improved accuracy, KANG’s design inherently provides insights into the model’s decision-making process, eliminating the need for post-hoc explainability techniques. This paper discusses the methodology, performance, and interpretability of KANG, highlighting its potential for applications in domains where interpretability is crucial.

Primary Area: learning on graphs and other geometries & topologies

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.

Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Submission Number: 9757