Skip to main content

Abstract

Artificial intelligence (AI) is everywhere around us and it is reforming just about every aspect of technology and everyday life. Everything is smart nowadays, from phones to fridges and washing machines, the edge of wireless networks is intelligent. This is not surprising taking into consideration that intelligent edge is corroborating AI applications. As the result of new trends and new demands on the market, a new research direction name edge learning originated. Edge learning intersects wireless networking and machine learning in favor of overcoming limitations and restrictions that prevail in those. The main challenge in wireless networks, in particular the IoT sensor networks, is computational power. Consequently, one of the main focuses of edge learning is how to overcome the limited computing power, along with limited data, one of the main challenges in machine learning. To be able to overcome mentioned restraints, massive data is distributed over a large number of edge devices, and leveraging the mobile edge computing platform is used. However, this union is not without faults. The main question is, how to obtain reliable and secure communication between the edge server and devices? In the same way, a strategy for transforming traditional machine learning algorithms into algorithms for learning from distributed data is another critical aspect. This talk will address these research challenges, security and privacy issues, and related existing solutions.

Search