Date and Time: 22 January 2021, 11am EST (New York/Toronto)
- Connection Details will be provided prior to the seminar.
- Please RSVP for connection details: https://forms.gle/C2vx9NexAg9QnmU36
Title: Artificial Intelligence for Packet Networks: The Cognitive Packet Network
Speaker: Erol Gelenbe, FIEEE FACM FIFIP
Professor, Institute of Theoretical and Applied Informatics, Polish Academy of Sciences, Honorary Professor University of Electronic Science and Technology China, Researcher University Cote d’Azur, Nice & Sophia Antipolis, France
Brief Abstract: We present the Cognitive Packet Network (CPN), which is an AI based routing protocol that uses Quality of Service (QoS), Security and Energy Optimisation, as objectives within a Goal Function that needs to be minimized. Network measurements collected by Smart Packets (SPs) update Random Neural Network Oracles with Reinforcement Learning, to dynamically select network paths so that end users can convey their payload traffic with a performance that matches the Goal as closely as possible. The Goal includes conventional QoS metrics such as delay and loss, as well as Real-Time objectives, as well as Energy Consumption and Security. Payload traffic is forwarded using source or segment routing, selected through the Reinforcement Learning, and SPs conduct their exploration using a node by node process to seek the best path. CPN has been implemented in various contexts: on 10-40 node test-beds, on an intercontinental scale as an overlay network, within SDN routers, and as a means to convey task requests over the Internet to Cloud servers. Our presentation will detail the CPN algorithm and the Random Neural Networks that are used to implement the Oracles. We will present related experimental measurements and results. The work has appeared in journals including Communications of the ACM, Proceedings of the IEEE, IEEE J. Sel. of Areas in Comms., and conferences including ICC and Globecom.
Short biography: A graduate of METU, Ankara, Turkey, Erol Gelenbe is listed in the Mathematical Genealogy Project of the American Mathematical Society among the 10 top PhD supervisors in the mathematical sciences. He is the inventor of the Random Neural Network and the eponymous G-Networks, the Cognitive Packet Nework routing algorithm, the Energy Packet Network paradigm, and diffusion approximations for network performance. His work won him election to Fellowship of the Royal Academy of Science of Belgium, the Science Academies of Hungary, Poland and Turkey, the National Academy of Engineering of France, and Academia Europaea, he is active in European research programs. He received the Mustafa Prize in 2017 (before Ugur Sahin’s same prize in 2019), as well as the ACM SIGMETRICS Life-Time Achievement Award and several other awards. He was made Knight of the Legion of Honour and Commander of Merit by the President of France, and Commander of Merit of the Republic of Italy by the President of Italy.