Machine Learning (ML) has been enjoying an unprecedented surge in applications that solve problems and enable automation in diverse domains. Primarily, this is due to the explosion in the availability of data, significant improvements in ML techniques, and advancement in computing capabilities. Undoubtedly, ML has been applied to various mundane and complex problems arising in network operation and management. There are various surveys on ML for specific areas in networking or for specific network technologies. This survey is original, since it jointly presents the application of diverse ML techniques in various key areas of networking across different network technologies. In this way, readers will benefit from a comprehensive discussion on the different learning paradigms and ML techniques applied to fundamental problems in networking, including traffic prediction, routing and classification, congestion control, resource and fault management, QoS and QoE management, and network security. Furthermore, this survey delineates the limitations, give insights, research challenges and future opportunities to advance ML in networking. Therefore, this is a timely contribution of the implications of ML for networking, that is pushing the barriers of autonomic network operation and management.

/pdf/a-comprehensive-survey-on-machine-learning-for-networking-3t5h3kbpe8.pdf

A comprehensive survey on machine learning for networking: evolution, applications and research opportunities

In this paper, a survey of the literature of the past 15 years involving machine learning (ML) algorithms applied to self-organizing cellular networks is performed. In order for future networks to overcome the current limitations and address the issues of current cellular systems, it is clear that more intelligence needs to be deployed so that a fully autonomous and flexible network can be enabled. This paper focuses on the learning perspective of self-organizing networks (SON) solutions and provides, not only an overview of the most common ML techniques encountered in cellular networks but also manages to classify each paper in terms of its learning solution, while also giving some examples. The authors also classify each paper in terms of its self-organizing use-case and discuss how each proposed solution performed. In addition, a comparison between the most commonly found ML algorithms in terms of certain SON metrics is performed and general guidelines on when to choose each ML algorithm for each SON function are proposed. Lastly, this paper also provides future research directions and new paradigms that the use of more robust and intelligent algorithms, together with data gathered by operators, can bring to the cellular networks domain and fully enable the concept of SON in the near future.

/pdf/a-survey-of-machine-learning-techniques-applied-to-self-5ggfpvrhwz.pdf

A Survey of Machine Learning Techniques Applied to Self-Organizing Cellular Networks

Trends in IoT based solutions for health care: Moving AI to the edge

An ensemble learning and fog-cloud architecture-driven cyber-attack detection framework for IoMT networks

With the advent of new technologies and the fast pace of human life, patients today require a sophisticated and advanced smart healthcare framework that is tailored to suit their individual health requirements. Along with 5G and state-of-the-art smart Internet of Things (IoT) sensors, edge computing provides intelligent, real-time healthcare solutions that satisfy energy consumption and latency criteria. Earlier surveys on smart healthcare systems were centered on cloud and fog computing architectures, security, and authentication, and the types of sensors and devices used in edge computing frameworks. They did not focus on the healthcare IoT applications deployed within edge computing architectures. The first purpose of this study is to analyze the existing and evolving edge computing architectures and techniques for smart healthcare and recognize the demands and challenges of different application scenarios. We examine edge intelligence that targets health data classification with the tracking and identification of vital signs using state-of-the-art deep learning techniques. This study also presents a comprehensive analysis of the use of cutting-edge artificial intelligence-based classification and prediction techniques employed for edge intelligence. Even with its many advantages, edge intelligence poses challenges related to computational complexity and security. To offer a higher quality of life to patients, potential research recommendations for improving edge computing services for healthcare are identified in this study. This study also offers a brief overview of the general usage of IoT solutions in edge platforms for medical treatment and healthcare.

/pdf/edge-intelligence-and-internet-of-things-in-healthcare-a-39gaf4b503.pdf

Edge Intelligence and Internet of Things in Healthcare: A Survey

Funding information National Science Foundation, Grant/Award Number: 1619346, 1559483, 1718956 and 1730650 Abstract Today, patients are demanding a newer and more sophisticated health care system, one that is more personalized and matches the speed of modern life. For the latency and energy efficiency requirements to be met for a real-time collection and analysis of health data, an edge computing environment is the answer, combined with 5G speeds and modern computing techniques. Previous health care surveys have focused on new fog architecture and sensor types, which leaves untouched the aspect of optimal computing techniques, such as encryption, authentication, and classification that are used on the devices deployed in an edge computing architecture. This paper aims first to survey the current and emerging edge computing architectures and techniques for health care applications, as well as to identify requirements and challenges of devices for various use cases. Edge computing application primarily focuses on the classification of health data involving vital sign monitoring and fall detection. Other low-latency applications perform specific symptom monitoring for diseases, such as gait abnormalities in Parkinson's disease patients. We also present our exhaustive review on edge computing data operations that include transmission, encryption, authentication, classification, reduction, and prediction. Even with these advantages, edge computing has some associated challenges, including requirements for sophisticated privacy and data reduction methods to allow comparable performance to their Cloud-based counterparts, but with lower computational complexity. Future research directions in edge computing for health care have been identified to offer a higher quality of life for users if addressed.

Edge computing in smart health care systems: Review, challenges, and research directions

Self-healing is a key desirable feature in emerging communication networks. While legacy self-healing mechanisms that are reactive in nature can minimize recovery time substantially, the recently conceived extremely low latency and high Quality of Experience (QoE) requirements call for self-healing mechanisms that are pro-active instead of reactive thereby enabling minimal recovery times. A corner stone in enabling proactive self-healing is predictive analytics of historical network failure logs (NFL). In current networks NFL data remains mostly dark, i.e., though they are stored but they are not exploited to their full potential. In this paper, we present a case study that investigates spatio-temporal trends in a large NFL database of a nationwide broadband operator. To discover hidden patterns in the data we leverage five different unsupervised pattern recognition and clustering along with density based outlier detection techniques namely: K-means clustering, Fuzzy C-means clustering, Local Outlier Factor, Local Outlier Probabilities and Kohonen's Self Organizing Maps. Results indicate that self-organizing maps with local outlier probabilities outperform K-means and Fuzzy C-means clustering in terms of sum of squared errors (SSE) and Davis Boulden index (DBI) values. Through an extensive data analysis leveraging a rich combination of the aforementioned techniques, we extract trends that can enable the operator to proactively tackle similar faults in future and improve QoE and recovery times and minimize operational costs, thereby paving the way towards proactive self-healing.

Enabling proactive self-healing by data mining network failure logs

In this article, we develop a statistical framework to quantify the area spectral efficiency (ASE) and the energy efficiency (EE) performance of a user-centric cloud based radio access network (UC-RAN) downlink. We propose a user-centric remote radio head (RRH) clustering mechanism, which: 1) provides significant improvement in the received signal-to-interference-ratio through selection diversity; 2) enables efficient interference protection by inducing repulsion among scheduled user-centric RRH clusters; and 3) can self-organize the cluster radius to deal with spatio–temporal variations in user densities. It is shown that under the proposed user-centric clustering mechanism, the ASE (bits/s/Hz/m2) maximizes at an optimal cluster size. It is observed that this cluster size is sensitive to changes in both RRH and user densities and, hence, must be adapted with variations in these parameters. Next, we formulate the cost paid for the UC-RAN capacity gains in terms of power consumption, which is then translated into the EE (bits/s/Joule) of the UC-RAN. It is observed that the cluster radius which maximizes the EE of the UC-RAN is relatively larger as compared with that which yields maximum ASE. Consequently, we notice that the tradeoff between the ASE and the EE of UC-RAN manifests itself in terms of cluster radius selection. Such tradeoff can be exploited by leveraging a simple two player cooperative game. Numerical results show that the optimal cluster radius obtained from the Nash bargaining solution of the modeled bargaining problem may be adjusted through an exponential weightage parameter that offers a mechanism to utilize the inherent ASE-EE tradeoff in a UC-RAN. Furthermore, in comparison with existing state-of-the-art non user-centric network models, our proposed scheme, by virtue of selective RRH activation and non overlapping user-centric RRH clusters, offers higher and adjustable system ASE and EE, particularly in dense deployment scenarios.

User-Centric Cloud RAN: An Analytical Framework for Optimizing Area Spectral and Energy Efficiency

With increasing cell density and the heterogeneity in the network, optimal user-cell association which is a well known open problem, will become an even more challenging issue Contrary to the current studies that address user-cell association problem for convectional HetNets with massive MIMO deployments in HF (high frequencies) ranges, in this paper we investigate user-cell association problem for dense two-tier networks with massive MIMO deployment both at macro and femto-tier operating in HF and mmWave spectrum, respectively We evaluate the performance of four user-cell association algorithms for massive MIMO deployment in a two-tier network under two different deployment scenarios: 1) HF-HF (both tiers operating in HF band); 2) HF-mmWave (MBSs operating in HF while FBSs in mmWave bands To this end, we model the association problem in form of a convex network utility maximization problem as a function of the downlink user throughput Contrary to the existing load aware association schemes that preclude the effect of bandwidth disparity in HF and mmWave bands, we propose a modified utility function that takes into account the effect of large bandwidth at mmWave bands The problem is solvable through centralized as well as distributed or user centric load aware user association schemes

What user-cell association algorithms will perform best in mmWave massive MIMO ultra-dense HetNets?

It is well known that current reactive network management would be unable to support the exponential increase in complexity and rapidity of change in future cellular networks. Keeping this in perspective, the goal of this paper is to investigate applicability of machine learning and predictive models to assess cell-level user quality of experience (QoE) in real-time. For this purpose, we leverage a 5 week LTE metrics data collected at cell level granularity for a national LTE network operator. Domain knowledge is applied to assess user QoE with network key performance indicators (KPIs), namely scheduled user throughput, inter-frequency handover success rate and intra-frequency handover success rate. Results indicate that applying boosted trees model on a subset of carefully selected non-collinear features allows high accuracy threshold-based estimation of user throughput and inter-frequency handover success rate. We also exploit the periodic nature of cell data characteristics and apply a recently developed time series prediction model known as PROPHET for future QoE estimation. By employing machine learning and data analytics on network data within an end-to-end framework, network operators can proactively identify low performance cell sites along with the influential factors that impact the cell performance. Based on the root cause analysis, appropriate corrective measures may then be taken for low performance cell sites.

Umair Sajid Hashmi

Papers

Edge computing in smart health care systems: Review, challenges, and research directions

Enabling proactive self-healing by data mining network failure logs

User-Centric Cloud RAN: An Analytical Framework for Optimizing Area Spectral and Energy Efficiency

What user-cell association algorithms will perform best in mmWave massive MIMO ultra-dense HetNets?

Towards Real-Time User QoE Assessment via Machine Learning on LTE Network Data