DataSciEn'2018 - Learning from Scientific Data in Energy

IFPEN / Rueil-Malmaison (France) - 16-17 January 2018

Agenda 

TUESDAY 16 JANUARY

9:00-9:15      Opening of the conference
                    T. Faney1, F. Chazal2, Scientific Correspondents of DataSciEn'2018 (1 IFPEN, 2 Inria)

                     Welcome Address
                     X. Longaygue, Scientific Division (IFPEN)

9:15-10:00   Plenary Lecture by O. Grisel (Inria)
                  Predictive Modeling in Python with scikit-learn

SESSION 1: FRAMEWORKS AND PLATFORMS

10:00-10:30  Data challenges with modularization and code submission
                    B. Kégl (CNRS)

10:30-10:40   Presentation of the Challenge

10:40-11:00   Coffee break

11:00-11:30    Accelerating deep learning with HPC architectures
                   
F. Courteille  (Nvidia)              

11:30-12:00   The graphical user interface of OpenTURNS, a UQ software in simulation
                   
M. Baudin (EDF)

12:00-12:30   ATOUT : Advanced Tools for Optimization and Uncertainty Treatment
                   
D. Sinoquet (IFPEN)

12:30-13:45    Lunch

13:45-14:30   Plenary Lecture by P. Gallinari (UPMC, LIP6)
                    Machine Learning to Deep Learning

SESSION 2: ADVANCES IN ALGORITHMS

14:30-14:55  Transfer Learning with Convolutional Networks
                    Y. Grandvalet (UTC)
14:55-15:20   Deep Learning for Physical Processes: Incorporating Prior Scientific Knowledge
                    E. de Bezenac (UPMC, LIP6)

15:20-15:45  Topological descriptors for geometric data
                   S. Oudot (INRIA)

15:45-16:05  Coffee break

16:05-16:30  Multivariate parametric regression under shape constraints
                  
F.Wahl (IFPEN)

16:30-16:55 Statistical models with large number of observations
                  
N. Durrande (PROWLER.io)

16:55-17:20  Warp 10, a uniform approach to time series storage and analytics
                   M. Herberts (Cityzen Data)

17:20-17:45  Prognostics and Health Management (PHM) for Schlumberger tools
                   
L. Baur (Schlumberger)
         

17:45-20:00  Lab1 : Scikit-learn        Lab2 : NVidia   
17:45-20:30 Cocktail Party

               
WEDNESDAY 17 JANUARY     

9:00-9:45   Plenary Lecture by M. Lutz (Total)

SESSION 3: INDUSTRIAL SESSION

9:45-10:15  Overview of recent scientific visualization results at Total EP
                  
M. Plainchault and B. Conche (Total)

10:15-10:30  DataIA Institute
                  
B. Thirion (Inria)

10:30-11:00  Coffee break

11:00-11:30  How data science is changing renewables asset management
                   M. Anderhalt (Données Brutes)

11:30-12:00  Predicting flooding in a distillation column: a data-driven modeling approach
                   N. Behara (Total)

12:00-13:30   Lunch

13:30-14:00  AGATHE : Predictive maintenance applied to HVAC equipment
                   C. Mallet
(Engie)
14:00-14:30  Data science in the non-supervized context of wind power
             
    
P. Poncet (Engie)

14:30-15:00  Calibration and validation of 0D-1D models for energy forecasting
                  
M. Keller (EDF)

15:00-15:30 Coffee break

15:30-16:00  Inline inspection and external inspection of oil and gas pipelines                 

                  J. Igersheim and L. Querella (Total)

16:00-16:30  Hydropower scheduling: learning handmade corrections
             
    
M. Huard (Univ. Paris Sud)

16:30           Closing address

 

Download the program here