-
Conference Room Auditorium Level -1
- Registration and welcome coffee from 07.30
- Start of the TUTORIAL 14.00
- Exhibition open from 09.00 to 19.30
- Seated Lunch: 12.30
- End of TUTORIAL 17.00
Plenary Session
MPLS + SDN + NFV world
AI net Conference
Level 0
AI net Conference
Auditorium Level -1
AI net Conference
Auditorium Level -1
AI net Conference
Auditorium Level -1
Imen Grida Ben Yahia, Expert in Future Networks, Orange Labs
Imen Grida Ben Yahia is currently with Orange Labs, France, as Orange Expert in Future Networks. She is leading a research project, on Autonomic & Cognitive Management for software networks. She received her PhD degree in Telecommunication Networks from Pierre et Marie Curie University in conjunction with Télécom SudParis in 2008. Her current research interests are autonomic and cognitive management for software and programmable networks that include artificial intelligence for SLA and fault management, knowledge and abstraction for management operations, intent- and policy-based management. As such, she contributed to several European research projects like Servery, FP7 UniverSelf and the H2020 CogNet and authored several scientific conference and journal papers in the field of autonomic and cognitive management. Imen is Chair of the IEEE Technical Committee on Network Intelligence (ongoing). She is currently TPC chair of Netsoft 2018 and also a member of several TPCs and conference organizing committees.
Jose Manuel Sanchez Vilchez, Research Engineer, Orange Labs
Jose Manuel Sanchez VILCHEZ is a research engineer in Orange Labs (France), specialized on data science for resilience management, root-cause analysis and reputation for softwarized networks SDN/NFV. He received a Master of Science in Telecommunication in September 2009 and the master of communication technologies, systems and networks Polytechnic University of Valencia (Spain) in September 2013. He received his PhD in computer networks by Telecom Sud Paris and UPMC (France) in July 2016 for the thesis entitled Multi-layer self-diagnosis of services over programmable networks José contributed to past collaborative projects such as ANR-REFLEXION (project devoted to resilience of SDN and NFV) and 5GPPP ENSURE (devoted to trust and security aspects of 5G). He authored and coauthored several papers in the domain of fault management, self-healing, security, and resiliency of SDN/NFV networks. France.
This tutorial aims to present a methodology with machine learning and deep learning techniques applied to Network Data. The goal is to share the various steps and challenges going from raw data transformation and visualization to model generation and application including model performance evaluation.
A basic forecasting for network timeseries with LSTM and a classification of network incident are presented to exemplify the different steps.
- Intro to time series and RNN LSTM
- Data pre-processing (Raw network data and their possible transformation (e.g. scaling, normalization, outlier removal, differentiation, etc.)
- Visualization and network data analysis (distribution, etc.) Network data forecasting with LSTM is presented to illustrate how important and complex is to configure a neural network
- A Network Incident classification problem on based on real network data solved with Random Forest
- Several variations of the hyper parameters of RF are shown to assess the impact on the classification performance
- Discussion on Feature engineering