Interpretable Graph Neural Networks for Tabular Data

TL;DR

This paper introduces IGNNet, an interpretable graph neural network for tabular data that matches state-of-the-art performance and provides explanations aligned with true feature contributions without extra computational cost.

Contribution

The paper presents IGNNet, a novel GNN-based model that is inherently interpretable and capable of accurately modeling feature interactions in tabular data.

Findings

IGNNet performs on par with XGBoost, Random Forests, and TabNet.

Explanations from IGNNet align with true Shapley values.

IGNNet offers interpretability without additional computational overhead.

Abstract

Data in tabular format is frequently occurring in real-world applications. Graph Neural Networks (GNNs) have recently been extended to effectively handle such data, allowing feature interactions to be captured through representation learning. However, these approaches essentially produce black-box models, in the form of deep neural networks, precluding users from following the logic behind the model predictions. We propose an approach, called IGNNet (Interpretable Graph Neural Network for tabular data), which constrains the learning algorithm to produce an interpretable model, where the model shows how the predictions are exactly computed from the original input features. A large-scale empirical investigation is presented, showing that IGNNet is performing on par with state-of-the-art machine-learning algorithms that target tabular data, including XGBoost, Random Forests, and TabNet. At the same time, the results show that the explanations obtained from IGNNet are aligned with the true Shapley values of the features without incurring any additional computational overhead.