Programme détaillé

• 08:30 - Accueil
• 09:00 - I. NEW RELEASES, NEW GEMS
• 09:00 - Scikit-Learn, the State of the Union

  Gaël Varoquaux, Inria
  

• 09:20 - Anomaly and Novelty Detection with scikit-learn

  Alexandre Gramfort, Telecom ParisTech
  

• 09:50 - Gradient Boosting: a practical guide

  Olivier Grisel, Inria
  

• 10:20 - Pause
• 10:49 - II. LEARNING WITH SCIKIT-LEARN
• 10:50 - Se former par l'open source

  Tom Dupré la Tour
  

  Tout ce que Scikit-Learn peut apporter en parallèle d'un cursus académique : contribution à un projet open source, expérience collaborative, accélération de l'apprentissage par la revue par les pairs, etc.

• 11:20 - Bénéfices de scikit-learn pour l'enseignant-chercheur en machine learning

  Chloé-Agathe Azencott, Mines ParisTech
  

  Chargée de recherche au Centre de bioinformatique de Mines ParisTech et enseignant le machine learning et la bioinformatique, Chloé-Agathe Azencott partagera son expérience d'utilisatrice de scikit-learn.

• 11:40 - Enseigner le machine learning avec Scikit-Learn: L'expérience de Télécom ParisTech

  Alexandre Gramfort, Telecom ParisTech
  

• 12:00 - Déjeuner networking
• 13:29 - III. ANIMER LA COMMUNAUTE ET CONTRIBUER
• 13:30 - Comment contribuer à Scikit-Learn ?

  Loic Esteve, Inria
  

• 14:00 - Software Carpentry: teaching as a community effort

  Bartosz Telenczuk, CNRS Gif-sur-Yvette
  

  Since 1998, Software Carpentry has been teaching researchers in science, engineering, medicine, and related disciplines the computing skills they need to get more done in less time and with less pain. We'll be talking about the teaching (what and how) and the community effort to share the work and keep the workshops going.

• 14:15 - Rapid Analytics and Model Prototyping (RAMP): a collaborative challenge

  Balász Kégl, CNRS-In2p3

  We will be describing the RAMP, a rapid data challenge format and tool we developed at the Paris-Saclay Center for Data Science. A RAMP is, at the same time, a collaborative prototyping challenge, a training session for novice data scientist, a networking opportunity, and a social science observatory.

• 14:30 - Pause
• 15:14 - IV. SUCCESS STORIES & APPLICATIONS
• 15:15 - Contribution d'un industriel : Scikit Learn à l'échelle

  Jean-Paul Smets, Nexedi
  

• 15:30 - Import sklearn : practical way we use it, and future thoughts

  Florian Douetteau, Dataiku
  

  In this presentation, we  will  drill through several practical use cases for Scikit-Learn, in logistics or customer analytics. We will present full workflows that combine Python / R / Hadoop and Scikit-Learn to achieve practical business goals, and present some thoughts about how the ML ecosystem is moving forward  to better support those

• 16:00 - Software Engineers, the new Data Scientists

  Christophe Bourguignat, Zelros
  

  Cards are reshuffled, as machine learning usages are growing, specialised software libraries are becoming a commodity, and MOOCs bring knowledge to the mass. In this talk, we will see how modern applications need building blocks like scikit-learn, and a new breed of software engineers proficient in machine learning.

• 16:30 - Malware Classification with scikit-learn & open questions (QA, governance)

  Fabien Mangeant, Vincent Feuillard, Pierre Benjamin, Airbus
  

  Using Scikit-Learn's text-mining functionalities to build malware classification algorithms. Further questions about the project's maintenance, contribution, roadmap governance will be raised.
  

• 16:45 - OPEN DISCUSSION