Fighting Fire with Fire: Survey of Strategies for Counteracting The Complexity of Future Networks Management

The evolution of telecommunications systems towards complex 5G networks implies changes on many levels. In this paper, we concentrate on the management landscape of current and future networks, unfolding a vision of advanced analytics and machine learning approaches to mitigate the unavoidable complexity of future networks. We explore proficient Deep Learning mechanisms and Cognitive Frameworks in the context of future management system needs, looking at techniques that can deliver automated feature engineering, integrating more knowledge into the system and paving the road towards heavily data driven and automated decision making. With this perspective, we analyse techniques such as deep learning, active learning, hybrid learning, and their relevance to the future telecommunications systems.