Graph Style Transfer for Counterfactual Explainability

2025/05/23 by Bardh Prenkaj, Prenkaj, Bardh, Efstratios Zaradoukas +3
Computer Science · #Adversarial Robustness in Machine Learning #Explainable Artificial Intelligence (XAI) #FOS: Computer and information sciences #Machine Learning (cs.LG)

paper · pdf · doi:10.48550/arxiv.2505.17542

openalex publication_date 2025/05/23 · openalex created_date 2025/10/20 · openalex updated_date 2026/07/28

Abstract

Counterfactual explainability seeks to uncover model decisions by identifying minimal changes to the input that alter the predicted outcome. This task becomes particularly challenging for graph data due to preserving structural integrity and semantic meaning. Unlike prior approaches that rely on forward perturbation mechanisms, we introduce Graph Inverse Style Transfer (GIST), the first framework to re-imagine graph counterfactual generation as a backtracking process, leveraging spectral style transfer. By aligning the global structure with the original input spectrum and preserving local content faithfulness, GIST produces valid counterfactuals as interpolations between the input style and counterfactual content. Tested on 8 binary and multi-class graph classification benchmarks, GIST achieves a remarkable +7.6% improvement in the validity of produced counterfactuals and significant gains (+45.5%) in faithfully explaining the true class distribution. Additionally, GIST's backtracking mechanism effectively mitigates overshooting the underlying predictor's decision boundary, minimizing the spectral differences between the input and the counterfactuals. These results challenge traditional forward perturbation methods, offering a novel perspective that advances graph explainability.

Citations

• Towards Non-Adversarial Algorithmic Recourse
• Robust Stochastic Graph Generator for Counterfactual Explanations
• Multimodal Motion Conditioned Diffusion Model for Skeleton-based Video Anomaly Detection
• Are we certain it's anomalous?
• Global Counterfactual Explainer for Graph Neural Networks
• Robust Counterfactual Explanations on Graph Neural Networks
• Counterfactual Graphs for Explainable Classification of Brain Networks
• Graph Domain Adaptation: A Generative View
• MEG: Generating Molecular Counterfactual Explanations for Deep Graph\n Networks
• CF-GNNExplainer: Counterfactual Explanations for Graph Neural Networks
• Explainability in Graph Neural Networks: A Taxonomic Survey
• Masked Label Prediction: Unified Message Passing Model for Semi-Supervised Classification
• Universal Graph Transformer Self-Attention Networks
• GNNExplainer: Generating Explanations for Graph Neural Networks
• A Survey of Methods for Explaining Black Box Models
• Style Transfer in Text: Exploration and Evaluation
• Controlling Linguistic Style Aspects in Neural Language Generation
• Shakespearizing Modern Language Using Copy-Enriched Sequence-to-Sequence\n Models
• Is writing style predictive of scientific fraud?
• Style Transfer from Non-Parallel Text by Cross-Alignment
• Demystifying Neural Style Transfer
• Training Sparse Neural Networks
• Preserving Color in Neural Artistic Style Transfer
• Pointer Networks
• Propagation Kernels
• The Graph Neural Network Model

Related

• CLEAR: Generative Counterfactual Explanations on Graphs
• Adapting to Change: Robust Counterfactual Explanations in Dynamic Data Landscapes
• Graph Diffusion Counterfactual Explanation
• Robust Counterfactual Explanations on Graph Neural Networks
• Robust Stochastic Graph Generator for Counterfactual Explanations