Pour vous authentifier, privilégiez eduGAIN / To authenticate, prefer eduGAINeu

Open Software Initiative (OSI) day

Europe/Paris
Alexandre Gramfort (Telecom ParisTech, CNRS), Balázs Kégl (LAL)
Description

Day dedicated to data science tools and engineering efforts at CDS, with presentations from funded projects, students, and invited speakers.

The event will take place in the main Auditorium of the Linear Accelerator Laboratory (LAL), building 200 on the UPSud Orsay campus. Information on getting to LAL is available here.

Webcast of the event here.

Current program:
https://docs.google.com/document/d/1W_4VOXuXg15YJTMZC7ax20SMpbBJALdh0HZA1foZIEA/edit?usp=sharing

Come back home.

Participants
  • Albert Thomas
  • Alessandro Ferreira Leite
  • Alexandre Abadie
  • Alexandre Boucaud
  • Alexandre Gramfort
  • Amir Sani
  • Anatole Chessel
  • Andreas Mueller
  • Balázs Kégl
  • Basile STARYNKEVITCH
  • Boualam HASNOUN
  • Cedric Viou
  • Christophe Pradal
  • Cécile Germain
  • David Froger
  • Dialekti VALSAMOU
  • Djalel Benbouzid
  • Eric Buchlin
  • Estelle Chaix
  • Eva D'hondt
  • Evangelia Zacharaki
  • Fabian Pedregosa
  • Fernando Perez
  • Franck Bardol
  • Gael Varoquaux
  • Grigori Rybkin
  • Grégory VIGUIER
  • Guillaume Carbajal
  • Guillaume Chevrot
  • Guillaume Wisniewski
  • Hervé BREDIN
  • Irina Poltavchenko
  • Jiali Mei
  • Julien Cochennec
  • Karima Rafes
  • Karin DASSAS
  • Konrad Hinsen
  • Konstantin Petrov
  • Kyle Penner
  • Leonardo Campillos
  • Lorenzo De Santis
  • Loïc Estève
  • Mainak Jas
  • Marc Evrard
  • Marc Nicolas
  • Martin Perez-Guevara
  • mehdi cherti
  • Moussab Djerrab
  • Nicolas Aunai
  • Nicolas Goix
  • Olivier Grisel
  • Oscar Nájera
  • Paolo Crisafulli
  • Patrick Brockmann
  • Pratheeban Elancheliyan
  • Raphaël Nedellec
  • Robin Vogel
  • Samuel Lelièvre
  • Sarah COHEN-BOULAKIA
  • simon tournier
  • Sébastien Treguer
  • Thiago DOS SANTOS GUZELLA
  • Thomas Schmitt
  • Tom Dupré la Tour
  • Vladislav Bastrikov
  • Yannig Goude
  • Yousra Bekhti
    • 08:30 09:00
      Welcome participants & coffee 30m
    • 09:00 09:15
      Introduction 15m
      Orateur: Alexandre Gramfort (Telecom ParisTech, CNRS)
      Transparents
      Video
    • 09:15 12:30
      Session 1
      • 09:15
        Keynote Speech 1: Jupyter at BIDS: building the tools and the institutions for data science 1h
        In this talk, I will discuss both the architecture and recent developments in the Jupyter project, (the multi-language platform evolved from the IPython system for interactive Python). Jupyter is a project aimed at a broad class of problems in data science, and at UC Berkeley, its development today takes place in a new kind of institutional context: the Berkeley Institute for Data Science (BIDS). I will also discuss the genesis of BIDS, and how this interdisciplinary space tries to foster a culture of collaboration, open research, and where specifically, the development of open source software for scientific research is a first-class citizen.
        Orateur: Fernando Perez (UC Berkeley - original author of IPython and driver of Project Jupiter)
        Slides
        Video
      • 10:30
        Coffee Break 30m
      • 11:00
        Py-Earth, Multivariate Adaptive Regression Splines in Python 15m
        Orateur: Mehdi Cherti (CNRS/LAL, AppStat)
        Slides
        Video
      • 11:15
        MNE from shell scripts and Unix commands to Python 15m
        Orateur: Lorenzo de Santis
        Slides
        Video
      • 11:30
        Reproducible research in Bioinformatics: Scientific Workflows, provenance and beyond 30m
        Orateur: Sarah Cohen-Boulakia (U. Paris-Sud/LRI)
        Slides
        Video
      • 12:00
        LoOPS Network for developers at Paris-Saclay 15m
        LoOPS is a network of scientific software developers in Higher Education and Research Institutions in the southwest of Paris. The mission of the network is to facilitate the sharing of practices, expertise and knowledge among developers in order to allow everyone to progress in its software development activity.
        Orateur: Karin Dassas (CNRS/IDOC)
        Slides
        Video
      • 12:15
        The io.datascience DaaS platform 15m
        Orateur: Mme Karima Rafes (LRI)
        Video
    • 12:30 14:00
      Lunch Break 1h 30m
    • 14:00 17:30
      Session 2
      • 14:00
        Building a service-oriented platform for online physiological data analysis. 20m
        Nowadays it is possible to online collect, classify, and process data with the proper software platforms. Along with advances in signal processing and "big data" algorithms we can use them as research facilities or even as automatic diagnostic tools. Specifically, we are building such a platform applied to physiological signals. It involves a large group of researchers in mathematics, software engineers, and neurologists in the framework of the COGNAC-C project.
        Orateur: Miguel Colom (ENS-Cachan/CMLA)
        Slides
        Video
      • 14:20
        Automation for chemical data analysis techniques 20m
        Chemical data acquired from different analytical techniques and disparate analytics platforms has become high volume and lacks of well organized. The raw data is formed under several different proprietary formats. All data requires to convert to a common open data formats for storage, analysis, pre-treatment and treatment. The current scientific workflows are manually done and requires a certain knowledge of programming languages. We propose a solution which can technologize the scientific workflow. Furthermore, the system enables to automate workflows with less efforts and without having manually initiate the complex configuration.
        Orateur: Diem Bui Thi
        Slides
        Video
      • 14:40
        Information Extraction Challenge for Gene Regulation Network in plant 20m
        Orateur: Estelle Chaix (INRA/MIG)
        Slides
        Video
      • 15:00
        Coffee Break 30m
      • 15:30
        Mining brain imaging data: lessons learned from nilearn and joblib 20m
        This talk will present useful patterns and lessons learned for efficient application of machine learning to data too big to fit in memory, in the context of brain imaging applications. The focus here is sharing insights gained while using and developing nilearn and joblib, as many of the lessons are not specific to brain imaging.
        Orateur: Loïc Estève (INRIA)
        Slides
        Video
      • 15:50
        A customizable framework for neurophysiology data management and provenance tracking 20m
        Orateur: Jonathan Duperrier (CNRS)
        Slides
        Video
      • 16:10
        Anomaly detection algorithms in Scikit-Learn 15m
        Orateur: Nicolas Goix (Telecom Paristech)
        Slides
        Video
      • 16:25
        Structured Prediction with Opera-Lib 15m
        Orateur: Romain Brault (CNRS/IBISC)
        Slides
        Video
      • 16:40
        Keynote Speech 2: Software and engineering efforts at NYU Center for Datascience 50m
        Orateur: Andreas Mueller (NYU)
        Slides
        Video