Sites Inria

Version française

Statistical sciences

Magalie Quet (*) - 30/05/2018

Statistics in the spotlight at Saclay during the 50th Statistics Days

© Fotolia TTStudio

This year, the fiftieth edition of the Statistics Days is being held from Monday 28 May to Friday 1 June.The EDF Lab Paris-Saclay will host nearly 450 French and international scientists, present 16 lectures and 63 parallel sessions (including 18 special sessions), clocking up nearly 100 hours of statistics in five days!This year's edition, under the scientific leadership of Francis Bach, researcher in the Sierra team at Inria's Paris centre, is organised by Christian Derquenne, senior researcher at EDF R&D, and Pascal Massart, professor at the Université Paris-Sud.They tell us more about this event, dedicated entirely to statistics.

A snapshot of statistics in France

The Journées de Statistique have been organised every year since 1970 by the Société Française de Statistique (SFdS). They are the most important scientific event in the French-speaking statistics community , with each edition drawing over 400 researchers, teachers and practitioners. Since their inception, the Journées de Statistique have been held a number of times in France (Marseille 2010, Toulouse 2014, Rennes 2014, Lille 2015, Montpellier 2016 and Avignon 2017), as well as in Belgium (Brussels 1982, 1992 and 2012, Brussels and Louvain-la-Neuve 2002), in Tunisia (Tunis 2011), Canada (Quebec 1996, Ottawa 2008), Morocco (Fez 2000) and Switzerland (Lausanne 1987, Neuchâtel 1994).

This year's edition, the 50th, shines a spotlight on the Saclay Plateau and its flagship scientific facilities, since the EDF Lab Paris-Saclay* has the privilege of hosting the event. "This edition, like the ones before it, is a snapshot of the state of play in statistics in France . And with its very international line-up of guest speakers, it is also a window on the world.Like a symbol of current advances in the discipline, the scientific programme opens with the Lucien Le Cam Lecture, which showcases the work of Arkadi Nemirovski, eminent professor from the Georgia Institute of Technology's College of Engineering, in statistics and in optimisation, two fields that are now closely related, thanks to the growing importance of big data processing," says Pascal Massart, vice-chairman of the organising committee.

Topics reflect current scientific issues

This year, the topics were chosen for their direct bearing on current scientific issues and in particular the rise of artificial intelligence , which places data at the centre of scientific concerns. Francis Bach, a prominent scientist in statistics and machine learning, confirms this: "As chairman of the scientific committee, I gathered a dozen leading figures from the broader scientific community to form the scientific committee.It was this committee that picked the guest speakers, all leading scientists in statistics orrelated fields such as machine learning or optimisation, for example. This year, we wanted to highlightthree topics with a strong scientific impact: "data science and learning"; "statistical challenges in industry";and lastly "energy and the environment". This is also reflected in the choice of guest speakers and special sessions."

Christian Derquenne, chairman of the organising committee, tells us more about the choice of topics:

  • "The first topic,"data science and learning" , is perfectly in keeping with current developments in statistics, in that big data processing and blends of existing and new techniques are helping to solve extremely complex problems.The data scientist's job is to choose the right tools in any given situation, with all of the contingencies that that entails.We can finally start dreaming of discovering theory through experimentation.
  • "Statistical challenges in industry" are the natural outcome of these new approaches and of the way certain scientific problems are currently being solved.But we may be wrong and, as researchers, it is also in our genes to challenge our hypotheses, models and results.This second topic obviously also ties in with the fact that the event is being held at an industrial research facility.
  • And lastly "energy and the environment" was chosen for a purpose:apart from the fact that the event is taking place at an energy utility's premises, it represents two indissociable aspects of sustainable development."

Keywords: Science Statistique Machine learning Optimisation Apprentissage Industrie Énergie Inria