Abstract
In this webinar, we will present a closed-loop automation system making use of AI to automatically mitigate anomalous states of the connect-compute software infrastructure. We will describe different functional blocks of the algorithmic framework, and in particular the anomaly detection module using federated learning, the AI function scheduling to meeting detection performance targets while mitigating AIF stragglers, the data-pipeline system design to ensure high accuracy, real-time data preprocessing and arrival , and a reinforcement learning framework to automatically find reconfiguration intents. The webinar will end with a preliminary demonstration of key system blocks and will highlight further research in this area.
The webinar will be introduced by Stefano Secci, head of the ROC team, and will be run live from Cnam. Remote participants can register from this MS Teams link.
Location: room 17.1.14, 292 rue St Martin, Cnam, Paris, France.
Speakers
Nour YELLAS (https://www.linkedin.com/in/nour-el-houda-yellas-0b177a174)
Nour-El-Houda Yellas is a Ph.D. candidate in Cnam, Paris, and teaching assistant at Sorbonne University. She received the master’s degree in network and engineering from Paris Saclay University in 2019. Her Doctoral research concerns mainly the automation and optimization of 5G mobile networks orchestration in multiaccess edge computing infrastructures using data analytics.
Naresh MODINA
Naresh Modina is a Post doctoral candidate at Cnam, he joined ROC team at Cnam (Paris), France on 1st January 2023. He received the M.Sc. degree in telecommunications engineering from Politecnico di Milano, Italy in 2019. He received the PhD in computer science from Avignon University, France in 2022. His current line of research is focused on applying AI in Wireless networks.
Patient NTUMBA
Patient NTUMBA-WA-NTUMBA is a post-doctoral researcher at Cedric laboratory of Cnam (Paris) in the ROC team. He joined the ROC team in February 2023 where he is carrying out research on data pipelining systems for in-network learning. He received the master’s degree in distributed systems and networks from Université de Franche-Comté (France) in 2017. In 2022, he received the PhD in computer science from Sorbonne University and carried out his doctoral research at INRIA Paris. His research focuses on the Internet of things, data streams, networks, distributed systems, and optimization.
Chi-Dung Phung
Chi Dung PHUNG (chi-dung.phung) is a research engineer at CNAM. Previously, he held research engineering positions at LIP6, UPMC, from 2013 to 2018, and at Orange Lab in 2019. He earned his Ph.D. from UPMC in 2018. His research interests include Internet control protocol design and experimentation.
Salah Bin-Ruba
Salah Bin-Ruba is a research engineer at CEDRIC lab of CNAM, Paris. He received his M.Sc. in Data Science and Network intelligence from Telecom SudParis in 2020. He is currently working on applying machine learning for anomaly detection in 5G infrastructure. His main research focus is on automation and auto configuration of networks using AI.
