Login

Net-Zero Machine Learning for Anomaly Detection in 6G Networks

BA, MA, MP
State: Open
Published: 2025-02-17

As 6G networks move toward commercial deployment, they promise ultra-high data rates, low latency, and intelligent automation. However, these networks are also expected to be sustainable and aligned with net-zero carbon emission goals. Traditional machine learning (ML) models used for network management and security often consume significant energy, making them unsuitable for sustainable AI applications. Anomaly detection is a critical security and performance-monitoring task in 6G networks, identifying cyberattacks, network failures, and performance bottlenecks. However, existing deep learning-based anomaly detection methods are energy-intensive and unsuitable for resource-constrained environments like edge computing and IoT-based 6G deployments. This project proposes a Net-Zero Machine Learning (NZML) approach that enables energy-efficient anomaly detection in 6G while maintaining high detection accuracy. Current anomaly detection methods in 6G networks rely on complex deep learning models that:

This research aims to develop a lightweight and energy-efficient ML framework for anomaly detection that ensures low power consumption while maintaining high detection performance.

This study aims to:

 

  1. Develop a Net-Zero Machine Learning (NZML) framework for anomaly detection in 6G networks.
  2. Optimize AI models using techniques such as pruning, quantization, and knowledge distillation to reduce power consumption.
  3. Integrate energy-efficient Federated Learning (FL) for anomaly detection across multiple 6G nodes.
  4. Leverage edge computing to perform real-time anomaly detection with minimal carbon footprint.
  5. Evaluate and benchmark the sustainability of ML models in terms of energy efficiency, detection accuracy, and response time.

Some Sources to consider:

1.      1. Li, Baolin, Siddharth Samsi, Vijay Gadepally, and Devesh Tiwari. "Clover: Toward sustainable ai with carbon-aware machine learning inference service." In Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis, pp. 1-15. 2023.

2.       2. Sthankiya, Kishan, Nagham Saeed, Greg McSorley, Mona Jaber, and Richard G. Clegg. "A Survey on AI-driven Energy Optimisation in Terrestrial Next Generation Radio Access Networks." IEEE Access (2024).

3.       3. Saafi, Salwa, Olga Vikhrova, Gábor Fodor, Jiri Hosek, and Sergey Andreev. "AI-aided integrated terrestrial and non-terrestrial 6G solutions for sustainable maritime networking." IEEE Network 36, no. 3 (2022): 183-190.

4.       4. Zhang, Peng, Yong Xiao, Yingyu Li, Xiaohu Ge, Guangming Shi, and Yang Yang. "Towards net-zero carbon emissions in network AI for 6G and beyond." IEEE Communications Magazine (2023).

5.     5. Benzaïd, Chafika, Fahim Muhtasim Hossain, Tarik Taleb, Pedro Merino Gómez, and Michael Dieudonne. "A federated continual learning framework for sustainable network anomaly detection in o-ran." In 2024 IEEE Wireless Communications and Networking Conference (WCNC), pp. 1-6. IEEE, 2024.

6.     6. Mata, Luís, Marco Sousa, Pedro Vieira, Maria Paula Queluz, and António Rodrigues. "Optimising Energy and Spectral Efficiency in Mobile Networks: A Comprehensive Energy Sustainability Framework for Network Operators." IEEE Access (2025).

20% literature review, 20% design, 50% implementation, 10% documentation

Supervisors: Chao Feng, Nasim Nezhadsistani

back to the main page