Is Meta-Path Attention an Explanation? Evidence of Alignment and Decoupling in Heterogeneous GNNs

2026-02-09

AI Summary

This work investigates whether meta-path attention in heterogeneous graph neural networks genuinely reflects the actual contribution of meta-paths and under what conditions it becomes decoupled from their true influence. To this end, the authors propose MetaXplain, a meta-path-aware explanation framework that requires no model modification and integrates gradient-based, perturbation-based, and Shapley-value-based explainers through view decomposition, pattern-compliant perturbations, and fusion-aware attribution for post-hoc analysis. The study presents the first systematic evidence that meta-path attention can both align with and significantly diverge from explanation outcomes, introducing the MP-AEA metric to quantify this consistency. Experiments demonstrate that MetaXplain outperforms random baselines, and retraining on explanation-derived subgraphs effectively denoises inputs, yielding performance improvements or preservation for HAN and HAN-GCN on ACM, DBLP, and IMDB benchmarks.

Technology Category

• Humans and AI: Explainable AI (XAI) for Human Understanding
• Machine Learning: Graph-based Machine Learning
• Search and Optimization: Metareasoning and Metaheuristics

Application Category

• Graph Algorithms and Modeling for the Web: Algorithms and analysis for heterogeneous, signed, attributed, multi-relational, temporal, higher-order, and annotated Web-related graphs
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Meta-path-based heterogeneous graph neural networks aggregate over meta-path-induced views, and their semantic-level attention over meta-path channels is widely used as a narrative for ``which semantics matter.''We study this assumption empirically by asking: when does meta-path attention reflect meta-path importance, and when can it decouple? A key challenge is that most post-hoc GNN explainers are designed for homogeneous graphs, and naive adaptations to heterogeneous neighborhoods can mix semantics and confound perturbations. To enable a controlled empirical analysis, we introduce MetaXplain, a meta-path-aware post-hoc explanation protocol that applies existing explainers in the native meta-path view domain via (i) view-factorized explanations, (ii) schema-valid channel-wise perturbations, and (iii) fusion-aware attribution, without modifying the underlying predictor. We benchmark representative gradient-, perturbation-, and Shapley-style explainers on ACM, DBLP, and IMDB with HAN and HAN-GCN, comparing against xPath and type-matched random baselines under standard faithfulness metrics. To quantify attention reliability, we propose Meta-Path Attention--Explanation Alignment (MP-AEA), which measures rank correlation between learned attention weights and explanation-derived meta-path contribution scores across random runs. Our results show that meta-path-aware explanations typically outperform random controls, while MP-AEA reveals both high-alignment and statistically significant decoupling regimes depending on the dataset and backbone; moreover, retraining on explanation-induced subgraphs often preserves, and in some noisy regimes improves, predictive performance, suggesting an explanation-as-denoising effect.

Problem

• meta-path attention
• heterogeneous GNNs
• explanation alignment
• attention decoupling
• GNN explainability

Innovation

• MetaXplain
• heterogeneous GNNs
• meta-path attention
• post-hoc explanation
• MP-AEA