Challenges 2018
During this second session, 7 challenges from the Challenge data website were presented :
- " Disease prediction from the genome " presented by Gilles Wainrib from Owkinen in collaboration with Inserm. The aim is to predict whether an individual will develop a certain disease based on data specific to his or her genome.
- " Predicting wind power production " presented by Paul Poncet of Engie. The aim is to predict the wind power output of four wind turbines based on their various operating parameters.
- " Predicting the dynamics of supercooled liquids presented by François Landes from Giulio Biroli 's team, IPhT CEA & ENS Paris. The aim is to predict the mobility of different particles within a system from their initial positions and velocities.
- " Sleep stage classification " presented by Valentin Thorey from Rythm. The aim is to predict an individual's sleep stage from 30 seconds of EEG, accelerometer and pulse oximeter signals.
- " Prediction of residual oil saturation presented by Jean-François Lecomte of IFPEN. The aim is to use the 3D pore structure of a rock to determine its type, as well as to predict the quantity of trapped residual oil (crude oil) and the size distribution of residual oil clusters.
- " Predicting the winning team of an NBA game " presented by Sébastien Loustau from LumenAI. The aim is to predict which team will win an NBA game, based on key indicators (score, fouls, etc.) of how the first half unfolded, recorded every second.
- " Publication Approval Prediction " presented by Nies Lubbers from Dassault Systèmes. The aim is to predict for each user of a social and collaborative application the 5 publications they are most likely to like, based on a history of publication approval by a set of users.
