Machine Learning Paradigms for Next-Generation Wireless Networks

Abstract

Next-generation wireless networks are expected to support extremely high data rates and radically new applications, which require a new wireless radio technology paradigm. The challenge is that of assisting the radio in intelligent adaptive learning and decision making, so that the diverse requirements of next-generation wireless networks can be satisfied. Machine learning is one of the most promising artificial intelligence tools, conceived to support smart radio terminals. Future smart 5G mobile terminals are expected to autonomously access the most meritorious spectral bands with the aid of sophisticated spectral efficiency learning and inference, in order to control the transmission power, while relying on energy efficiency learning/inference and simultaneously adjusting the transmission protocols with the aid of quality of service learning/inference. Hence we briefly review the rudimentary concepts of machine learning and propose their employment in the compelling applications of 5G networks, including cognitive radios, massive MIMOs, femto/small cells, heterogeneous networks, smart grid, energy harvesting, device-todevice communications, and so on. Our goal is to assist the readers in refining the motivation, problem formulation, and methodology of powerful machine learning algorithms in the context of future networks in order to tap into hitherto unexplored applications and services.

Terahertz communication: The opportunities of wireless technology beyond 5G

“…Actually, ML can be widely used in modeling a number of technical problems including large scale MIMOs, D2D networks as well as HetNets [34]. In particular, as THz communication consists of access points in either ubiquitous WiFi networks or base-station clustering, unsupervised learning strategies are required.…”

Softwarization of Mobile Network Functions towards Agile and Energy Efficient 5G Architectures: A Survey

“…Machine learning (ML) techniques constitute a promising solution for network management and energy savings in cellular networks. According to [66,67], they can be categorized as supervised, unsupervised, or reinforcement learning-based. Supervised and unsupervised learning, respectively, indicate whether the samples from the dataset are labelled or not.…”

“…This is pretty common in practical cases and, as such, it may eventually become the preferred learning technology for many application domains. The basic concept of ML algorithms and the corresponding applications according to the category of supervised, unsupervised, and reinforcement learning is presented in [66]. ML can be used in modeling various technical problems for next-generation systems.…”

“…RL and DRL can be used for inferring mobile users policies under unknown network conditions, for example, BS association under unknown ESI of the BS, in EH networks. The reader is referred to [66,67] for a comprehensive review of ML techniques applied to cellular networks.…”

A machine learning-based framework for preventing video freezes in HTTP adaptive streaming

“…Despite its simplicity, this method has proven to be very effective in many real classification problems. As an example, Jian et al report how a K-NN classifier can be used to optimize the behavior of 5G networks, in terms of handover selection or energy savings [25]. The training time for the different classifiers is in the order of tens of minutes on a Dell Latitude E5530 running Ubuntu 12.04 LTS 64-bit, Intel Core i5-3320M CPU @ 2.60GHz processor and 8 GB of memory.…”