Exakis Nelite, part of Magellan Partners group, the first Microsoft pure-player partner in France, announced in early January its appointment to the “Data Lake for Nuclear” project as part of the French Government’s France Relance program. Exakis Nelite has formed a consortium with Heverett Group, Nuclear Valley, Limos and Aquilla, whose objective is to enable French nuclear players to share their “dormant data” and collaborate on training AI algorithms. Funding will be provided by BpiFrance and the Research Tax Credit (CIR).

In September 2020, the Government launched an exceptional €100 billion economic stimulus plan to quickly and sustainably turn around the French economy. 470 million and has allocated €200 million to investment support funds. This is how the Data Lake for Nuclear project was able to be formed, supported by EDF and GIFEN, the French nuclear industry’s professional union with more than 200 members, “to apply AI techniques to benefit the quality and performance of nuclear achievements.”

Pooling “dormant data”

Like many companies in other sectors, companies in the nuclear sector may resort to data collection. Exploited for a single use for a specific control, this stored data then becomes dormant data. The ambition of this project is to pool this data in order to use it in the training of algorithms and the objective of Exakis Nelite, the creation of a cloud platform in compliance with the security and anonymization requirements associated with the nuclear world. Didier Zetoun, president of Exakis Nelite, explains:

“

This project allows Exakis Nelite’s teams to implement a shared and secure cloud architecture to serve the digitalization needs of the French nuclear industry.”
Naturally, the Microsoft Cloud Azure platform will be used for this implementation; it will also serve as the technological foundation for using its specific features in the areas of Data, artificial intelligence, governance and security. The technological barriers to be overcome will concern the use of homomorphic encryption for data, the partitioning and sealing of artificial intelligence environments as well as the design of a Data Lake capable of addressing the heterogeneity of data. According to Exakis Nelite:

“Once deployed, this project will be a real gas pedal but also an incubator for all projects in the industry.”

Translated from Exakis Nelite lauréat du projet «Data lake for Nuclear» du Plan de Relance