What question did this study set out to answer?

This research aims to enhance the security of Graph Neural Networks against backdoor attacks using a novel framework.

July 9, 2026Electronics0 citationsOpen Access

An Explainable CS-Mitigation Triangular (ECSMT) Framework to Secure Graph Neural Networks

SESabah EttahriDominion University College SESergio Pallas EnguitaOld Dominion University CCChung‐Hao ChenUniversity of Manchester

Key Points

This research aims to enhance the security of Graph Neural Networks against backdoor attacks using a novel framework.
Proposed the Explainable Complex System-Mitigation Triangular Framework (ECSMT) integrating robust training, graph regularization, and data sanitization.
Conducted empirical evaluations across three benchmark domains: AIDS, MUTAG, and PROTEINS using a Graph Isomorphism Network (GIN) backbone.
Analyzed defensive performance under a baseline 5% backdoor subgraph trigger injection ratio.
ECSMT achieved a Clean Accuracy (CA) of 97.33% (±0.62%) while reducing the Attack Success Rate (ASR) from 97.00% to 69.45% on the AIDS benchmark.
Cross-domain testing showed that small-scale datasets like MUTAG suffer from persistent trigger concentration, while PROTEINS exhibit high topological noise.
Implemented mapping of technical outcomes into an enterprise framework resulting in a 61% expenditure compression and a 98.93% reduction in systemic loss.

Abstract

This research addresses cyber risk by defending against backdoor attacks on Graph Neural Networks (GNNs). We propose the Explainable Complex System-Mitigation Triangular (ECSMT) Framework, which integrates Robust Training, Graph Regularization, and Data Sanitization into a lightweight, hardware-efficient defense layer. To evaluate structural generalizability, we conducted empirical evaluations across three distinct benchmark domains (AIDS, MUTAG, and PROTEINS) using a Graph Isomorphism Network (GIN) backbone. Under a baseline 5% backdoor subgraph trigger injection ratio, ECSMT achieves excellent utility retention, securing a Clean Accuracy (CA) of 97.33% (±0.62%) while reducing the Attack Success Rate (ASR) from 97.00% down to 69.45% on the primary AIDS benchmark. Cross-domain testing reveals that defensive efficacy is strongly constrained by dataset characteristics: small-scale datasets such as MUTAG suffer from persistent trigger concentration, while complex graph manifolds such as PROTEINS exhibit high levels of topological noise. Furthermore, mapping these technical outcomes into an enterprise asset framework yields a 61% expenditure compression at critical technological feeder locations and a 98.93% reduction in total systemic loss. This study indicates that the proposed triangular mitigation strategy offers a valuable, scalable blueprint for enhancing the technical resilience and prognostic economic modeling of critical infrastructure networks.

问 AI

Bookmark

View Full Paper