DIAMOND : Databases, Artificial Intelligence, Modeling and Optimization tools for the Digital Design of materials

Responsible : François WILLAIME, CEA Gif-sur-Yvette

Keywords: Platforms, codes, data, artificial intelligence, workflows

The main objective of this targeted project is to establish a digital infrastructure to accelerate materials development. It is based on the one hand on a platform dedicated to simulation codes at different scales and workflows for the automation of series of calculations, including high-throughput calculations and the processing of results by artificial intelligence (AI) and on the other hand on a database infrastructure coming from both experiments and simulations.

For each of these infrastructures, use cases from current targeted projects have been identified to serve as an example and ensure suitability with user needs; they concern the MOFsLEARNING, ESRF and FastNano projects. A demonstrator using these two infrastructures will also be developed, it will be dedicated to the development by machine learning of interatomic potentials for simulations at the atomic scale. The specifications for these infrastructures and this demonstrator will be established based on working groups with representatives of the main laboratories concerned. The codes and workflows platform will be developed and deployed at the GRICAD meso-computing center and the database infrastructure will be deployed at TGCC.