Please, enable JavaScript in your browser :)

data sharing without revealing your data || ActInSpace Challenges


Be a NewSpace Player #Space4.0

Data sharing without revealing your data !

Use space data with AI models through Federated Learning to improve performances




Raw space data is more and more accessible...

However large parts of space data (raw, processed) is not freely available for various reasons (commercial, security...)

The idea is to use segregated data in order to train AI models through Machine Learning.
Therefore, participants could benefit from improved models (algorithms) with collective data, without having to disclose their confidential data to others.


This challenge is to demonstrate that a service could be set up (for example through Federated Learning) using different sources of open and not open space data (raw or processed).

Identify data sources (for example: optical, SAR, Hyperpectral, IR, navigation, communication...) with possible applications and their respective constraints (considering sources compatibility, access to data and time constraints ...)

Carry out a market study ... and imagine a way ahead ...

Nota: "Federated Learning is defined as a distributed machine learning framework authorizing a model to be built from data distributed across different data owners without data sharing."


Discover the winners of the locals editions !
Posted on November 24, 2020
AIS webinar !
Posted on October 30, 2020
Only a few weeks until ActInSpace 2020 !
Posted on October 29, 2020
Logo ActInSpace
Home page
Logo ActInSpace
Logo ActInSpace
Logo ActInSpace
News / Blog
Logo ActInSpace
World map
Logo ActInSpace
Legal notice
Logo ActInSpace
Logo of AIRBUS
Logo of Invest In Toulouse
Logo of CLS
Logo of EBAN
Logo of AirZeroG
Logo of GSA
Logo of Suez
Logo of Continental
Logo of ESSP
Logo of KINEIS
Logo of WEkEO
Logo of Dassault Aviation
Logo of Gen Space
Logo of SES