FlowX: Towards Explainable Graph Neural Networks via Message Flows

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Gui, Shurui, Yuan, Hao, Wang, Jie, Lao, Qicheng, Li, Kang, Ji, Shuiwang
Format: Preprint
Published: 2022
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916077035847680
author Gui, Shurui
Yuan, Hao
Wang, Jie
Lao, Qicheng
Li, Kang
Ji, Shuiwang
author_facet Gui, Shurui
Yuan, Hao
Wang, Jie
Lao, Qicheng
Li, Kang
Ji, Shuiwang
contents We investigate the explainability of graph neural networks (GNNs) as a step toward elucidating their working mechanisms. While most current methods focus on explaining graph nodes, edges, or features, we argue that, as the inherent functional mechanism of GNNs, message flows are more natural for performing explainability. To this end, we propose a novel method here, known as FlowX, to explain GNNs by identifying important message flows. To quantify the importance of flows, we propose to follow the philosophy of Shapley values from cooperative game theory. To tackle the complexity of computing all coalitions' marginal contributions, we propose a flow sampling scheme to compute Shapley value approximations as initial assessments of further training. We then propose an information-controlled learning algorithm to train flow scores toward diverse explanation targets: necessary or sufficient explanations. Experimental studies on both synthetic and real-world datasets demonstrate that our proposed FlowX and its variants lead to improved explainability of GNNs. The code is available at https://github.com/divelab/DIG.
format Preprint
id arxiv_https___arxiv_org_abs_2206_12987
institution arXiv
publishDate 2022
record_format arxiv
spellingShingle FlowX: Towards Explainable Graph Neural Networks via Message Flows
Gui, Shurui
Yuan, Hao
Wang, Jie
Lao, Qicheng
Li, Kang
Ji, Shuiwang
Machine Learning
Artificial Intelligence
We investigate the explainability of graph neural networks (GNNs) as a step toward elucidating their working mechanisms. While most current methods focus on explaining graph nodes, edges, or features, we argue that, as the inherent functional mechanism of GNNs, message flows are more natural for performing explainability. To this end, we propose a novel method here, known as FlowX, to explain GNNs by identifying important message flows. To quantify the importance of flows, we propose to follow the philosophy of Shapley values from cooperative game theory. To tackle the complexity of computing all coalitions' marginal contributions, we propose a flow sampling scheme to compute Shapley value approximations as initial assessments of further training. We then propose an information-controlled learning algorithm to train flow scores toward diverse explanation targets: necessary or sufficient explanations. Experimental studies on both synthetic and real-world datasets demonstrate that our proposed FlowX and its variants lead to improved explainability of GNNs. The code is available at https://github.com/divelab/DIG.
title FlowX: Towards Explainable Graph Neural Networks via Message Flows
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2206.12987