New technologies are emerging for the Future Internet, and while networks are growing faster, they are also becoming more complex and difficult to manage. Consequently, network administrators have increasing limitations to operate, troubleshoot and optimize their networks. In this context, the network community (manufacturers, operators, researchers, etc.) is looking at Artificial Intelligence / Machine Learning (AI/ML) methods often with inflated expectations. Unfortunately, for many network applications, like network monitoring and troubleshooting, an AI/ML model is of little use if it cannot be interpreted by a human operator.

This project aims to develop an interpretable machine learning approach based in the Multivariate Big Data Analysis (MBDA) methodology for network applications. We will make use of a general framework called Networkmetrics, referring to the use of multivariate analysis and other interpretable machine learning tools in network applications. It stands for the combination of the application domain (network engineering) and the suffix “-metrics”, inherited from other disciplines where interpretable multivariate analysis has been widely adopted, both in academia and industry.

This project aims to make the next step forward for networkmetrics, with three interrelated goals:

  • GOAL 1 : To apply networkmetrics to technologies of the Future Internet. In particular, we would like to develop new approaches and algorithms suited for Software Defined Networking (SDN)

  • GOAL 2: To extend cutting-edge multivariate techniques to Big Data, in applications like fault detection and cybersecurity, traffic classification and network optimization.

  • GOAL 3: To promote networkmetrics in the international community in industry and academia.