Some useful information to use and play with the platform!


To use this service well, please, take a time to read the information below :

“Challenge Data” is an initiative supported by the CFM-ENS chair “Modèles et Sciences des Données”, to bridge the gap between real industrial problems, research and teaching in data science. It promotes a free exchange of data and algorithmic knowledge, for the development of education and research in data science on one hand, and industrial applications on the other.

The Challenge Data platform hosts machine learning challenges, based on data provided by large groups, start-ups or research labs. They are addressed to students, data scientists and researchers in machine learning and statistics. Each challenge is a supervised classification or regression problem, organized as a competition according to a Kaggle-like procedure. Challenges proposed typically cover a wide range of type of data (images, time series, text, etc.), domains (finance, medicine, energy, physics, etc.) and machine learning problems.

Each project includes a description of its context, its objectives and provides a supervised dataset for training. A participant submits the result of his algorithm on a test set and gets a score and a ranking according to a specified metric. A benchmark score is also provided to guide participants to the type of performance they should attain to be in the game. In order to avoid overfitting, submissions are limited to 2 every 24 hours and the test set in split equally between a public part on which public ranking is established and a private part on which final ranking and intermediary rankings are computed.

Participants to a challenge are invited to upload a final report describing their algorithm at the end of a challenge which can be made public or available only to challenge providers. Students registered to a course can on top submit a report at any time during the course which will be available to professors, and can elect to make it as well public or available to challenge providers, fostering the free exchange principle of this initiative.

Professors who register their class can specify a subset of projects for their students, follow their activity and rankings, and automatically download their reports.

Challenge providers can monitor activity on their challenge through a private dashboard and can also access live public and private rankings as well as reports communicated by participants.

For further information to register your class or submit a challenge, contact challenge.data@ens.fr.