Experiences in operational networks.


  • Registration and welcome coffee from 07.30
  • Start of the Conference 08.30
  • Exhibition open from 08.00 to 19.00
  • Seated Lunch: 12.30


Morning Chairman
Jean-Marc Uzé,
VP EMEA, Augtera Networks
Smart Telemetry/Data Quality Session
How Many Data do we Need to Build a Good Classifier for Network Management?
Discussing how we can apply AI techniques to improve IP/MPLS Networks operations and user experience. Applying few-shot learning, meta-learning, and a generative algorithm to different domains of network management; eNodeB data analysis for quality assurance in mobile network, packet and flow data analysis for Web service identification of encrypted transport connections.

Kohei Shiomoto, Professor, Tokyo City University

Kohei Shiomoto is a Professor, Tokyo City University, where he is engaged in Research and Education in Network Management for the Internet, Mobile, and Cloud. Before joining Tokyo City University as Full Professor, he spent 28 years in NTT laboratories and was engaged in advanced research on SDN/NFV and Network Analytics, Communication Traffic & Service Quality, Traffic Engineering, IP and Optical Networking, (network architecture, IETF standardization, multi-layer traffic engineering), architecture design for high-speed IP/MPLS label switching router, IP router design, routing algorithm, and IETF GMPLS standardization, development of commercial ATM switching systems, research of high-speed networking technologies including ATM traffic control, measurement-based admission control, and ATM switch design.   From 1996 to 1997, he was on leave from NTT to join Department of Computer Science and Engineering, Washington University in St. Louis, MO, USA as Visiting Researcher and was engaged in research of high-speed networking technologies including dynamic flow switching and soft-state ATM switching.   He received the B.E., M.E., and Ph.D degrees in information and computer sciences from Osaka University, Osaka in 1987 1989, and 1998, respectively.

Intent Based Analytics
Talking about IBA - Intent Based Analytics (smart telemetry) - Analytics that are defined and derived from user intent and used to validate the intended state and are immediately actionable.
This is as opposite to “dump telemetry”, trying to get as much data as possible and then doing nothing with it, since it is too late and lacks context (un-actionable data).

Jeff Tantsura, Head of Networking Strategy, Apstra
ML/AI to Collect and Correlate Data from your Network and the Internet
In order to get a full perspective on one’s network, one must also keep in mind the ‘bigger picture’ of the internet and correlate its big data with the big data from their own network. Examining the idea behind matching and correlating these data sets and advanced ML/AI techniques that can be used to automate data collection and correlation.

Craig Labovitz, CTO, Deepfield Business Unit, Nokia
Reinforcement Learning Session
Reinforcement Learning for Telecommunication Network: from Opportunistic Spectrum Access to IoTs
In reinforcement learning, an agent chooses actions in order to maximize the rewards given by a dynamic environment. As the environment is initially unknown, the agent has to interact with it to gather information. Moreover, only the reward of the chosen actions is revealed. That is why the agent faces the exploration/exploitation dilemma: he has to explore loosely estimated actions in order to build a better estimate, and he would like to maximize his cumulated reward by playing the empirically best actions.

Raphaël Féraud, Research Engineer, Orange Labs

Raphaël Féraud received his Ph.D from the university of Rennes in 1997. As a research engineer of Orange, he has worked on various use cases of Artificial Intelligence: image processing, and specifically on face detection, dynamic allocation of resources for telecommunication networks, targeting of Internet advertising, churn, fraud detection, marketing optimization…  He led research projects on machine learning and Big Data. Since 2012, his research focuses on online learning, reinforcement learning and multi-armed bandits.

Coffee Break / Exhibition / Interop Showcase

A Deep Learning System to Accurately Diagnose the Wireline Access Networks
Discussing deep learning developments to offer an accurate, proactive and network-wide solution to troubleshoot the wireline access networks. Indeed, thanks to various innovative aspects, the system “learns” the key patterns of defects from routine measurements of the line and pro-actively suggests the cause of the impairment to the technician.

Nicolas Dupuis, Senior Machine Learning Scientist, AI Strategy and Innovation Lead, DMTS, Nokia

Winner of the two last Global AI Conferences organized by Nokia Bell Labs, Distinguished Member of Technical Staff and main author of 30+ patents in the field of Broadband communication & AI, Nicolas Dupuis is a Tech. Lead for AI Innovation & Machine Learning developments within Nokia.   His strong technical background and his long-term contact with the market led him to create pioneering product features globally deployed at major service providers.

Deployment Scenarios to Augment Network Operations with AI
Addressing deployment challenges that are facing customers when exploring and implementing AI-based solutions for they operations, as well as appropriate approaches. Introducing various options as experienced with customers in the last 18 months in order to better anticipate future deployments.

Prof. Mérouane Debbah, IEEE Fellow, CentraleSupélec, Huawei
11.30 Panel
AIOps: How to Start
Amy Cameron, Senior Analyst, STL Partners

Amy works with world-leading operators and tech companies to define opportunities and strategies in the B2B and AI fields. She leads STL Partners’ research into application of AI in telecoms, as well as the Growing Enterprise Revenues research stream.   Before joining STL in 2017, she was Head of ICT Research at Fitch Solutions.   Amy speaks French and Mandarin and holds an MSc in Chinese Politics from the School of Oriental and African Studies.

Seated Lunch


Afternoon Chairman
Jean-Marc Uzé,
VP EMEA, Augtera Networks
Self Healing Networks Session
What is Self Healing Enterprise and how to Become One?
Talking about how self-healing enterprise, level of self-healing and how autonomous IT enables self-healing.
Business benefits of self-healing are elevated digital user experience, optimization of IT resources, and achieve uninterrupted operations for the enterprises.

Paddy Padmanabhan, Founder & CEO, Appnomic Systems
Self-healing Networks: Software Defined Transport Network Automation
Network resources can be efficiently used by steering traffic to meet low latency, high bandwidth and resilience requirements. Highlighting the tenets of Software Defined Transport Network Automation enabling:
  • Service-oriented transport provisioning
  • Intent-driven real-time transport network optimization

Ranga Maddipudi, Product Line Manager, Cisco
Coffee Break / Exhibition / Interop Showcase

Intent-based Intelligence : Systematic Self-healing Network
To support service agility, the network must evolve from a static resource system to a flexible and dynamic system that can meet business objectives. With the global network status control, a closed-loop network can be automatically built and maintained based on human service intentions to implement intelligent and systematic self-healing networks.

Hu Junjie, Chief of Transport Network Marketing Dept., ZTE Corporation
AI Chips and Intelligent Lossless Algorithms
Enabling zero packet loss on the Ethernet, unleashing the full computing potential of AI, and realizing the integration of computation, storage, and communication on DCNs:
  • Congestion-caused packet loss on traditional Ethernet becoming the bottleneck for improving computing power in the AI era: a packet loss rate of 0.1% resulting in a decrease of 50% in computing power
  • AI-based intelligent plane for DCNs, realizing zero packet loss

Xiaofeng Yang, Data Center Network Senior Architect, Huawei

Mr.Yang has 15-year experience in telecom carrier network, in charge of the research ,innovation, standard promotion work of data center network.   He and his team are taking an active part in the data center network solution innovation, esp in innovative lossless network and its related applicable scenarios research including AI computing cluster and distributed storage, SDN automation innovation and applicable network practice and on-going promotion, etc.

End of Conference Day Two Track 3