





















Abstract:The trust in the predictions of Graph Neural Networks is limited by their opaque reasoning process. Prior methods have tried to explain graph networks via concept-based explanations extracted from the latent representations obtained after message passing. However, these explanations fall short of explaining the message passing process itself. To this aim, we propose the Concept Graph Convolution, the first graph convolution designed to operate on node-level concepts for improved interpretability. The proposed convolutional layer performs message passing on a combination of raw and concept representations using structural and attention-based edge weights. We also propose a pure variant of the convolution, only operating in the concept space. Our results show that the Concept Graph Convolution allows to obtain competitive task accuracy, while enabling an increased insight into the evolution of concepts across convolutional steps.
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2604.20082 [cs.LG] |
| (or arXiv:2604.20082v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2604.20082 arXiv-issued DOI via DataCite (pending registration) |
From: Lucie Charlotte Magister [view email]
[v1]
Wed, 22 Apr 2026 00:57:06 UTC (365 KB)
此内容由惯性聚合(RSS阅读器)自动聚合整理,仅供阅读参考。 原文来自 — 版权归原作者所有。