THIS WEBSITE IS THE OLD WEBSITE, IT IS NO MORE MAINTAINED, GO TO THE NEW ONE.

Our challenges

All the challenges offered by the platform since its creation.

Challenge Data 2018

16/11/2017 - 27/12/2018
Challenge Data 2016-2017

21/10/2016 - 01/01/2018
Challenge Data 2015-2016

21/10/2015 - 01/07/2016

Challenge Data 2018

Solve a major physics problem by finding how to predict the dynamics of glass-forming liquids — Solve a major physics problem by finding how to predict the dynamics of glass-forming liquids
What is the physical mechanism behind the slowing down of the dynamics in super-cooled liquids close to the glass transition? Being able to predict the future mobility of particles with high accuracy

Help ENGIE improve wind power production! — Help ENGIE improve wind power production!
Predict wind power production from wind turbine operational data. Video of the challenge's presentation at Collège de France is available at: https://www.college-de-france.fr/site/stephane-malla

Oze-Energies, improving energy consumptions with statistical learning — Oze-Energies, improving energy consumptions with statistical learning
This data challenge aims at introducing a new statistical model to predict energy consumptions and inside temperatures using observations stored in the Oze-Energies database. We propose to build a sim

Celebrity identification challenge — Celebrity identification challenge
In this challenge, we provide you with faces extracted from different videos. The goal is to recognize the people on screen given a small database of celebrities faces. Video of the challenge's pre

Prediction of Residual Oil Saturation — Prediction of Residual Oil Saturation
Use rock porous microstructure given as a 3D image to predict the amount of residual oil and the size of the oil clusters in this microstructure. Video of the challenge's presentation at Collège d

Disease prediction from DNA data — Disease prediction from DNA data
The goal of this project is to predict whether a given individual will have a specific disease (undisclosed for confidentiality reasons) or not, starting from the sequence of bases in the genome of t

Prediction of transaction claims status — Prediction of transaction claims status
The goal of this challenge is to predict if a transaction will be subject to a claim from the user (broken, not received, ...) or not. Results from this challenge will help us to improve the quality o

NBA challenge — NBA challenge
The main goal of the challenge is to predict the winning NBA team at the end of a game when we only have information about the first two quarters of the game. This is a binary classification problem (

Learning Sleep Stages from physiological signals on Dreem Headband — Learning Sleep Stages from physiological signals on Dreem Headband
We propose to use Dreem headband (dreem.com) to perform sleep stage scoring on 30 seconds epochs of biophysiological signals. Video of the challenge's presentation at Collège de France is availabl

Volatility prediction in financial markets — Volatility prediction in financial markets
Use past volatilities and price changes of financial instruments to predict future volatility and control the risk of financial portfolios Community forum for sharing ideas and making faster progre

Predict the expected answer! — Predict the expected answer!
Millions of questions are asked every year about drugs. There is a limited number of query types but the same question could be asked in many ways. Therefore, understanding what information people exp

Recommender Systems in the 3DEXPERIENCE Platform — Recommender Systems in the 3DEXPERIENCE Platform
The goal of the challenge is to build a recommender system for the social and collaborative application that is integrated within the 3DEXPERIENCE Platform (Dassault Systèmes). For each user, this sy

Predict the hourly production of all the French power units (157 units, with different technologies: nuclear, gas...) — Predict the hourly production of all the French power units (157 units, with different technologies: nuclear, gas...)
We already have a good model for the national production of these different technologies, but having a picture at the power unit level is interesting both on a scientific level, and on a more philosop

Challenge Data 2016-2017

Cluster actor faces from TV show — Cluster actor faces from TV show
In this challenge, we provide you with faces extracted from 20 episodes of a TV show. The goal is to gather, for each movie, all the faces that belong to the same actor into homogeneous and complete c

Prediction of products reviews interests — Prediction of products reviews interests
PriceMinister is one of the leading marketplace of E-commerce in France, and is part of the Rakuten group, leader in Japanese market. The goal of this challenge is to predict if a review of a user

Predict the Crude Oil production trend. — Predict the Crude Oil production trend.
Predict the Crude Oil production's trend base on the previous year Crude Oil data.

Prediction of trading activity within the order book — Prediction of trading activity within the order book
With past order book snapshots, try to predict if any trading activity will happen within the next second.

Failure risk prediction on pipeline network — Failure risk prediction on pipeline network
The goal of this challenge is to predict which pipes have the highest failure risk in the two next years.

Post Recommendation in a Collaborative Platform — Post Recommendation in a Collaborative Platform
The goal of the challenge is to build a recommender system for the social and collaborative application that is integrated within the 3DEXPERIENCE Platform (Dassault Systèmes). For each user, this sy

Age prediction from EEG signals — Age prediction from EEG signals
The goal of this challenge is to predict the age of subject from his brain activity during deep sleep and the sequence of his sleep stages over a night.

Oze Energies : Optimizing energy consumptions — Oze Energies : Optimizing energy consumptions
This data challenge aims at introducing a new statistical model to analyze energy consumptions in several buildings using observations received from sensors. These observations are decomposed into: -

Predict which clients reduced their consumption — Predict which clients reduced their consumption
EDF is constantly seeking to improve customer experience, notably in helping its customers to better understand and thus better manage their energy consumption.

Predict air quality at the street level — Predict air quality at the street level
We propose to predict the air quality level at the street level in several French cities. To reach this goal, we rely on an increasingly popular kind of model in the atmospheric science community call

Challenge Data 2015-2016

Retrieve all the products that are mentioned in magazines via a mobile application — Retrieve all the products that are mentioned in magazines via a mobile application
Selectionnist mobile application allows users to take pictures of products or pages that they see in magazines, and retrieve the products that were mentioned in this page, so that they can put them in

Classify bird songs — Classify bird songs
Can you guess which bird is singing ? Identify 50 bird species from all over South America by analyzing their songs, as recorded in the wild.

Electrod inversion detection in ECGs — Electrod inversion detection in ECGs
ECG is a temporal signal measuring the electric activity of the heart. It is very easy to record: 3 electrodes are placed at the extremities of the limbs, and 6 are placed on the chest. These 9 el

Using face recognition and clustering to associate actor faces in movies — Using face recognition and clustering to associate actor faces in movies
Given detected and tracked faces in movies, we want to gather in clusters faces that belong to the same identity.

Prediction of transaction volumes in financial markets — Prediction of transaction volumes in financial markets
Use past traded volumes of financial instruments to predict future traded quantities and optimize trading execution

Attentional Selection in a Cocktail Party — Attentional Selection in a Cocktail Party
Use electroencephalography (EEG) data to identify which speaker a person is paying attention to.

Machine Learning for Sensors Reduction in Body Posture Tracking — Machine Learning for Sensors Reduction in Body Posture Tracking
The body posture tracking can be done with 10 inertial sensors placed on the body. We want to explore a new approach where the tracking can be done with a smaller number of sensors (in this project on

SNCF – Predictive maintenance - Fault predictions on trains — SNCF – Predictive maintenance - Fault predictions on trains
SNCF propose a predictive maintenance challenge in partnership with Quantmetry. This is a classification challenge: you will have to predict faults on trains’ doors. An interesting aspect of this ch

Fetal Abdomen Localization — Fetal Abdomen Localization
Given a training set of fetal abdomen ultrasound images with ground truth bounding box of abdomen localization, you will have to train an object localization algorithm able to find the position of the

Predict pollution in an asian megacity — Predict pollution in an asian megacity
!!!! IMPORTANT !!!! !!!! IMPORTANT : files (input train) have been updated on the 10th of december in the evening !!!! !!!! IMPORTANT !!!! We want to predict the concentration of some pollutants

Sleep stages classification — Sleep stages classification
How to predict the sleep stage using electro-encephalogram (EEG) and accelerometer ?

Information