Events 2022

📍 Machine Learning & Industry Day (2022)

Overview

This event, co-organised by EDF R&D and the French ENBIS network
(frENBIS), is part of our ongoing efforts to promote statistics and machine learning methods within companies and industrial environments.

The morning session was dedicated to a training course on online expert aggregation methods, delivered by
Pierre Gaillard, researcher at INRIA Grenoble and author of the
OPERA – Online Prediction by Expert Aggregation package.

The theoretical foundations of expert aggregation were presented, together with practical illustrations in R and Python (notebooks).

The afternoon featured talks from industry, start-ups, and academia, focusing on concrete industrial applications of machine learning.

🧭 Practical Information

  • 📍 Venue: EDF Lab Saclay
  • 📅 Date: Monday, May 23, 2022
  • 💰 Registration: free of charge, mandatory registration (online form)

🕒 Programme

Morning – Training Session

Time Content
09:00 – 09:30 Welcome coffee
09:30 – 11:00 Expert aggregation – Lecture 1
11:00 – 11:15 Break
11:15 – 12:30 Expert aggregation – Lecture 2

Lunch break

Afternoon – Applied Talks

Time Title Speaker
13:45 – 14:25 AI in healthcare: navigating between algorithmic innovation and regulatory challenges Malo Huard, MILVUE
14:25 – 15:05 Data for Good: using mobile data to address societal challenges Stefania Rubrichi, Orange
15:25 – 16:05 Temperature prediction with expert aggregation Léo Pfitzner, Météo-France

📍 Satellite Event at JdS 2022 (Lyon)

Organisation

frENBIS (the French local network of ENBIS) and the
MALIA – MAchine Learning and Artificial Intelligence
group of the SFdS jointly organised a satellite event during the
Journées de la Statistique 2022 in Lyon.

🧭 Practical Information

  • 📅 Date: Friday, June 17, 2022
  • 🎯 Scientific theme:
    Statistical learning for time series data: new perspectives and industrial applications

Organising committee:
Christine Kéribin, Jean-Michel Poggi, Jairo Cugliari, Anne Gégout-Petit,
Paul Poncet, Bertrand Iooss, Yannig Goude

🎯 Objectives of the Day

The goal of this event was to present innovative statistical learning methods for time series data in an industrial context.

  • Morning: an introductory training session on deep learning for time series, including hands-on practical work.
  • Afternoon: a workshop consisting of three talks presenting original applications in telecommunications, air pollution, and energy.

A significant amount of time was dedicated to discussion, with the aim of fostering collaboration between: academics, industrial R&D teams, and start-ups.

🕒 Programme

Morning – Training (09:00 – 12:00)

  • Introduction to deep learning methods for time series data
  • Lecture and practical session (Python)
  • Instructor: Vincent Guigue (LIP6)

Afternoon – Workshop (14:00 – 16:30)

  • 14:00 – Introduction (SFdS / frENBIS)
  • 14:10 – 14:55 – Julien Jacques (Lyon 2)
    Co-clustering of multivariate functional data for air pollution analysis
  • 14:55 – 15:40 – Stefania Rubrichi (Orange)
    Using relay antenna data to study COVID propagation
  • 15:40 – 16:25 – Vincent Le Guen (EDF R&D)
    Deep learning for photovoltaic power forecasting from ground-based images
  • 16:30 – Conclusion

☕ frENBIS Virtual Coffee Break (2022)

frENBIS organised an online webinar designed to:

  • encourage informal exchanges,
  • collect feedback, expectations, and proposals from network members.

🧭 Practical Information

  • 📅 Date: Friday, April 8, 2022
  • ⏰ Time: 2:00 pm – 4:00 pm
  • 💻 Format: online

Programme

  • Machine learning for anomaly detection: methods, applications, and challenges
    Kim Phuc Tran (ENSAIT)
  • Graph clustering for intrusion detection
    Baptiste Gregorutti (LumenAI)
  • Real-life data analysis in Essilor R&D Innovation
    Aurélie Le Cain (Essilor)

A final discussion session allowed participants to synthesise expectations and proposals for the future development of the frENBIS network.