Generating Synthetic Datasets for Anomaly Detection in 5G Networks using AI/ML

The adoption of 5G networks introduces unprecedented performance capabilities, including ultra-low latency, high-speed data transfer, and massive device connectivity. However, these advancements also bring challenges in monitoring, securing, and maintaining network performance. Anomaly detection is critical for ensuring the reliability and security of 5G networks, but the scarcity of labeled datasets due to the complexity and novelty of the technology poses a significant hurdle. This thesis explores the development of synthetic datasets to address this limitation by leveraging Artificial Intelligence (AI) and Machine Learning (ML) techniques.

This thesis aims to design a framework for generating realistic and representative synthetic data for 5G network scenarios, simulating both normal and anomalous behaviors. By integrating techniques such as Generative Adversarial Networks (GANs), Variational Autoencoders (VAEs), and simulation-based approaches, the study will evaluate the quality, usability, and effectiveness of these datasets in training and testing anomaly detection models. Key metrics such as precision, recall, detection rate, and generalization ability will be analyzed. The thesis contributes to enhancing the applicability of AI/ML in 5G network monitoring, supporting advancements in automated network management and threat detection.

References:

[1] Lam, J. and Abbas, R. "Machine learning based anomaly detection for 5g networks." arXiv preprint arXiv:2003.03474 (2020). 

[2] Kim, Y.S., Kim, Y.E. and Kim, H. "A Model Training Method for DDoS Detection Using CTGAN under 5GC Traffic." Computer Systems Science & Engineering 47, no. 1 (2023).

[3] Samarakoon, S., Siriwardhana, Y., Porambage, P., Liyanage, M., Chang, S.Y., Kim, J., Kim, J. and Ylianttila, M. "5g-nidd: A comprehensive network intrusion detection dataset generated over 5g wireless network." arXiv preprint arXiv:2212.01298 (2022).

[4] Farzaneh, B., Shahriar, N., Al Muktadir, A.H., Towhid, M.S. and Khosravani, M.S. "DTL-5G: Deep transfer learning-based DDoS attack detection in 5G and beyond networks." Computer Communications 228 (2024): 107927.

[5] Wang, Z., Fok, K.W. and Thing, V.L. "Exploring Emerging Trends in 5G Malicious Traffic Analysis and Incremental Learning Intrusion Detection Strategies." arXiv preprint arXiv:2402.14353 (2024).

[6] Pandey, C., Tiwari, V., Rathore, R.S., Jhaveri, R.H., Roy, D.S. and Selvarajan, S. "Resource-efficient synthetic data generation for performance evaluation in mobile edge computing over 5g networks." IEEE Open Journal of the Communications Society 4 (2023): 1866-1878.

[7] Yin, Y., Lin, Z., Jin, M., Fanti, G. and Sekar, V. "Practical gan-based synthetic ip header trace generation using netshare." In Proceedings of the ACM SIGCOMM 2022 Conference, pp. 458-472. 2022.