As telecommunications networks become increasingly complex, the integration
of advanced technologies such as network digital twins and generative
artificial intelligence (AI) emerges as a pivotal solution to enhance network
operations and resilience. This paper explores the synergy between network
digital twins, which provide a dynamic virtual representation of physical
networks, and generative AI, particularly focusing on Generative Adversarial
Networks (GANs) and Variational Autoencoders (VAEs). We propose a novel
architectural framework that incorporates these technologies to significantly
improve predictive maintenance, network scenario simulation, and real-time
data-driven decision-making. Through extensive simulations, we demonstrate how
generative AI can enhance the accuracy and operational efficiency of network
digital twins, effectively handling real-world complexities such as
unpredictable traffic loads and network failures. The findings suggest that
this integration not only boosts the capability of digital twins in scenario
forecasting and anomaly detection but also facilitates a more adaptive and
intelligent network management system.

电信网络变得越来越复杂，与网络数字孪生和生成式人工智能（AI）等先进技术的整合成为增强网络运营和弹性的关键解决方案。本文探讨了网络数字孪生与生成式人工智能（尤其关注生成对抗网络（GANs）和变分自编码器（VAEs））之间的协同作用。我们提出了一个新颖的架构框架，将这些技术结合起来，显著提高了预测性维护、网络场景模拟和实时数据驱动决策的能力。通过广泛的模拟，我们证明了生成式人工智能如何提高网络数字孪生的准确性和运营效率，有效处理不可预测的流量负载和网络故障等现实世界的复杂性。研究结果表明，这种整合不仅增强了数字孪生在情景预测和异常检测方面的能力，还促进了更自适应、智能的网络管理系统。

将生成式人工智能与网络数字孪生集成，增强网络运营

Integrating Generative AI with Network Digital Twins for Enhanced  Network Operations

Fifth-generation (5G) core networks in network digital twins (NDTs) are
complex systems with numerous components, generating considerable data.
Analyzing these data can be challenging due to rare failure types, leading to
imbalanced classes in multiclass classification. To address this problem, we
propose a novel method of integrating a graph Fourier transform (GFT) into a
message-passing neural network (MPNN) designed for NDTs. This approach
transforms the data into a graph using the GFT to address class imbalance,
whereas the MPNN extracts features and models dependencies between network
components. This combined approach identifies failure types in real and
simulated NDT environments, demonstrating its potential for accurate failure
classification in 5G and beyond (B5G) networks. Moreover, the MPNN is adept at
learning complex local structures among neighbors in an end-to-end setting.
Extensive experiments have demonstrated that the proposed approach can identify
failure types in three multiclass domain datasets at multiple failure points in
real networks and NDT environments. The results demonstrate that the proposed
GFT-MPNN can accurately classify network failures in B5G networks, especially
when employed within NDTs to detect failure types.

提出了一种将图傅里叶变换与消息传递神经网络相结合的方法来准确分类第五代核心网络中的故障类型，并在网络数字孪生中展示其潜力。

基于图神经网络的网络数字孪生模型的 5G 网络故障分类

Beyond 5G Network Failure Classification for Network Digital Twin Using  Graph Neural Network

Network digital twins (NDTs) facilitate the estimation of key performance
indicators (KPIs) before physically implementing a network, thereby enabling
efficient optimization of the network configuration. In this paper, we propose
a learning-based NDT for network simulators. The proposed method offers a
holistic representation of information flow in a wireless network by
integrating node, edge, and path embeddings. Through this approach, the model
is trained to map the network configuration to KPIs in a single forward pass.
Hence, it offers a more efficient alternative to traditional simulation-based
methods, thus allowing for rapid experimentation and optimization. Our proposed
method has been extensively tested through comprehensive experimentation in
various scenarios, including wired and wireless networks. Results show that it
outperforms baseline learning models in terms of accuracy and robustness.
Moreover, our approach achieves comparable performance to simulators but with
significantly higher computational efficiency.

本文提出了一个基于学习的网络数字孪生用于网络模拟器，它可以综合表示无线网络中的信息流，并通过将网络配置映射到关键绩效指标来进行快速实验和优化。通过在多种情景下全面测试，结果表明该方法比传统的基于模拟的方法具有更高的精确性和鲁棒性，同时实现了比模拟器更高的计算效率。