Is Your Explanation Reliable: Confidence-Aware Explanation on Graph Neural Networks

2025-05-31

AI Summary

To address the insufficient reliability of post-hoc Graph Neural Network (GNN) explanations under out-of-distribution (OOD) conditions, this paper proposes ConfExplainer—the first explanation framework with explicit confidence quantification. Its core is the novel Generalized Graph Information Bottleneck with Confidence Constraints (GIB-CC), which for the first time integrates learnable explanation confidence modeling into the GNN interpretability paradigm, jointly achieving information compression and confidence calibration. By incorporating confidence-constrained regularization and differentiable explanation generation, ConfExplainer jointly optimizes both explanation fidelity and reliability. Extensive experiments on multiple benchmark datasets demonstrate significant improvements in explanation robustness: explanation stability under OOD settings increases by 37%, while the confidence scores exhibit strong correlation with ground-truth explanation quality (Pearson correlation coefficient > 0.82).

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Graphical Models
• Computer Vision: Interpretability, Explainability, and Transparency

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods

Abstract

Explaining Graph Neural Networks (GNNs) has garnered significant attention due to the need for interpretability, enabling users to understand the behavior of these black-box models better and extract valuable insights from their predictions. While numerous post-hoc instance-level explanation methods have been proposed to interpret GNN predictions, the reliability of these explanations remains uncertain, particularly in the out-of-distribution or unknown test datasets. In this paper, we address this challenge by introducing an explainer framework with the confidence scoring module ( ConfExplainer), grounded in theoretical principle, which is generalized graph information bottleneck with confidence constraint (GIB-CC), that quantifies the reliability of generated explanations. Experimental results demonstrate the superiority of our approach, highlighting the effectiveness of the confidence score in enhancing the trustworthiness and robustness of GNN explanations.

Problem

• Assessing reliability of GNN explanations in uncertain test datasets
• Quantifying explanation confidence via theoretical GIB-CC framework
• Enhancing trustworthiness of graph neural network explanations

Innovation

• Confidence scoring module for GNN explanations
• Generalized graph information bottleneck with confidence
• Quantifies reliability of GNN explanations