Hybrid spatial-temporal graph neural network Powered NDTs:Towards Next-Gen Smart Infrastructure Twins

2026-08-08

AI Summary

Existing network digital twins (NDTs) struggle to accurately predict the dynamics of complex mobile networks due to limited modeling capacity, hindering real-time optimization of intelligent infrastructure. To address this challenge, this work proposes a Hybrid Spatio-Temporal Graph Neural Network (HSTGNN), which innovatively integrates local neighborhood aggregation, spectral filtering, and a learnable attention mechanism into a multi-scale hybrid message-passing architecture, thereby overcoming the representational limitations of conventional single-branch GNNs. Experimental results demonstrate that HSTGNN significantly enhances both prediction accuracy and generalization capability, outperforming baseline models on key metrics: it achieves a coefficient of determination (R²) of 0.8816 (a 17.5% improvement), a mean absolute error (MAE) of 0.0300, and a root mean square error (RMSE) of 0.0458, effectively enabling high-fidelity network digital twins.

Abstract

Network Digital Twins (NDTs) enable proactive network management and optimization by predicting system behavior before control actions are applied to live infrastructures, supporting critical operations in Internet Service Provider (ISP) networks and wide-area networks (WANs). However, to anchor the superior performance NDTs promise to provide, key enabler techniques are required. Given that mobile networks are modeled as graphs, graph-based architectures such as graph neural networks (GNNs) have shown promising performance in modeling network behavior. This work proposes a novel Hybrid Spatial-Temporal Graph Neural Network (HSTGNN) architecture. Unlike single-branch GNN approaches, we propose a multi-scale design that combines three complementary message-passing paradigms: local neighborhood aggregation, spectral filtering, and learnable attention-based weighting. When benchmarked against other approaches, the proposed HSTGNN achieved superior performance delivering a coefficient of determination score of approximately 0.8816, 17.5\% better than the best baseline ChebNet. Furthermore, HSTGNN achieved the lowest Mean Absolute Error (MAE) of 0.0300, and Root Mean Squared Error (RMSE) of 0.0458, significantly outperforming baseline frameworks and certifying the proposed framework's capability in enabling NDTs.

Problem

• Network Digital Twins
• graph neural networks
• spatial-temporal modeling
• network behavior prediction
• mobile networks

Innovation

• Hybrid Spatial-Temporal Graph Neural Network
• Network Digital Twins
• Multi-scale Message Passing
• Graph Neural Networks
• Attention-based Weighting