A Survey of Machine Learning Algorithms for 6G Wireless Networks

TL;DR

This survey reviews machine learning techniques applicable to 6G wireless networks, highlighting their roles in reducing costs, enhancing performance, and enabling advanced features like network slicing and edge automation.

Contribution

It provides a comprehensive overview of ML algorithms for 6G, identifying open research problems and future directions in the field.

Findings

ML reduces power consumption and improves system performance.

ML enables high automation and dynamic network management.

Open research problems in ML for 6G are identified.

Abstract

The primary focus of Artificial Intelligence/Machine Learning (AI/ML) integration within the wireless technology is to reduce capital expenditures, optimize network performance, and build new revenue streams. Replacing traditional algorithms with deep learning AI techniques have dramatically reduced the power consumption and improved the system performance. Further, implementation of ML algorithms also enables the wireless network service providers to (i) offer high automation levels from distributed AI/ML architectures applicable at the network edge, (ii) implement application-based traffic steering across the access networks, (iii) enable dynamic network slicing for addressing different scenarios with varying quality of service requirements, and (iv) enable ubiquitous connectivity across the various 6G communication platforms. In this chapter, we review/survey the ML techniques which are applicable to the 6G wireless networks. and also list the open problems of research which require timely solutions.