Interpreting Manifolds and Graph Neural Embeddings from Internet of Things Traffic Flows

2026-02-05

AI Summary

This work addresses the limitations of existing network monitoring approaches in capturing dynamic relationships among IoT devices and the lack of interpretability in graph neural network (GNN) embeddings. To this end, we propose an interpretable analysis framework that integrates manifold learning with feature attribution. Our method is the first to map high-dimensional GNN embeddings onto a low-dimensional latent manifold, enabling both visualization of network state evolution and attribution of critical traffic features. The framework not only facilitates the identification of concept drift but also achieves an F1 score of 0.830 in intrusion detection tasks, effectively uncovering the structural dependencies and behavioral evolution mechanisms inherent in IoT traffic.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Data Mining & Knowledge Management: Anomaly/Outlier Detection
• Natural Language Processing: Interpretability, Analysis, and Evaluation of NLP Models

Application Category

• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Social Networks and Social Media: Social media analysis through the lenses of networks
• Web Mining and Content Analysis: Bridging structured and unstructured data

Abstract

The rapid expansion of Internet of Things (IoT) ecosystems has led to increasingly complex and heterogeneous network topologies. Traditional network monitoring and visualization tools rely on aggregated metrics or static representations, which fail to capture the evolving relationships and structural dependencies between devices. Although Graph Neural Networks (GNNs) offer a powerful way to learn from relational data, their internal representations often remain opaque and difficult to interpret for security-critical operations. Consequently, this work introduces an interpretable pipeline that generates directly visualizable low-dimensional representations by mapping high-dimensional embeddings onto a latent manifold. This projection enables the interpretable monitoring and interoperability of evolving network states, while integrated feature attribution techniques decode the specific characteristics shaping the manifold structure. The framework achieves a classification F1-score of 0.830 for intrusion detection while also highlighting phenomena such as concept drift. Ultimately, the presented approach bridges the gap between high-dimensional GNN embeddings and human-understandable network behavior, offering new insights for network administrators and security analysts.

Problem

• Interpretability
• Graph Neural Networks
• IoT Traffic
• Manifold Learning
• Network Visualization

Innovation

• interpretable GNN
• manifold visualization
• IoT traffic analysis
• feature attribution
• concept drift detection