A European strategy

CEA-List is dedicated to making its scientific ecosystem stronger across Europe through an ambitious partnership strategy and a clear positioning.

A particularly active stakeholder in research and innovation, CEA-List is stepping fully into its role, helping to represent the CEA in several key alliances and partnerships strategic to our areas of expertise. These include EFFRA for the factory of the future; ADRA for AI, data and robotics; BDVA for big data; ECSO for cybersecurity; INSIDE for smart digital systems; and Eclipse for open-source software. We are also part of five networks of excellence: AI4Media for AI for the media industry; TAILOR for trustworthy AI; HiPeac for computing systems; euRobin for reproducible AI; and dAIEDGE for trustworthy distributed AI.

In addition, CEA-List participates in or coordinates some 100 submissions to a wide range of calls for projects issued by Horizon Europe, Digital Europe, the European Defense Fund, Euratom, EuroHPC, and other EU Joint Undertakings like the Chips JU. The 94 European projects currently in progress address a variety of sectors encompassing industry, digital technology, healthcare, civil security, defense, agriculture, and the environment.
We are also tackling societal challenges around the circular economy through projects on digital product passports and the secure exchange of industrial data at the European level. These projects are, of course, aligned with our eco-innovation program.

Today, Europe needs world-class, high-impact technological research infrastructure where industrial companies of all sizes can test innovative AI-based concepts and products in conditions representative of their production environments. CEA-List leveraged the AI-MATTERS project (one of the European TEFs – Testing and Experimentation Facilities) to create PRISM. This unique facility, through the services it provides to SMEs, is helping make Europe’s manufacturing industry more competitive.

Eco-innovation

Environmental responsibility is one of CEA-List’s core values. We have been working on eco-innovation for several years now, and our efforts have matured into a research program specifically dedicated to eco-innovation. CEA-List researchers published an article detailing how we are helping to reduce the direct environmental and societal impacts of technology (SustainableTech)(1) and how we are using technology to help solve environmental and societal problems and assess indirect impacts (Tech4Sustainability), as well as the underlying methodologies based on existing standards. Our approach to eco-innovation is based on the conviction that both must be addressed simultaneously if we are to develop technologies backed by an assessment of both positive and negative environmental impacts to arrive at a net benefit.

Our researchers are gradually broadening the use of our eco-innovation process. One example is our research on Digital Product Passports (DPP); another is on software being developed for Life Cycle Assessments (LCA). In the area of waste from electrical and electronic equipment (WEEE), CEA-List technologies are being applied to the recycling of critical materials and the identification of undesirable materials in waste streams.

Figure 1 : Our eco-innovation process.

1 Sustainable technology and Technology for sustainability: The paths towards Eco-innovation, Chiara Sandionigi, Jean-François Berrée, Maxime Peralta, Ariane Piel, Bénédicte Robin, Electronics Goes Green 2024.

Major projects

The people at CEA-List, with their world-class scientific knowledge and deep understanding of industrial culture, have been pivotal in the emergence of ambitious projects to capitalize on and deploy CEA-List innovations.

@Istock
© P. Charlot / CEA

Here are a few examples:

  • The EU STARLIGHT project (2021-2025) built a sustainable, multidisciplinary European community around AI for security, bringing together researchers, representatives from law enforcement and industry, plus ethics experts, for a total of 50 partners from eighteen countries. Co-development techniques led to innovations deployed across eight test cases. A total of 50 interoperable AI tools for threat assessment, investigation, and cybersecurity were developed. Furthermore, the project established a robust ethical, legal, and regulatory compliance framework spanning data protection, traceability, and alignment with the AI Act.
  • The EU DeepGreen project (deepgreen.ai), backed by a consortium of over 20 partners, was established to build a sovereign platform for the deployment of embedded AI open to the entire European industrial ecosystem. The advances made under the DeepGreen project are being integrated into the Aidge open-source platform hosted by the Eclipse Foundation (eclipse.dev/aidge/). The project is helping restore industrial companies’ freedom to choose the technologies they use by supporting the development of AI solutions addressing a wide range of established and emerging markets.
  • A series of coordinated projects like EPI and MOSAICS-LP are focusing on the design of high-performance computing systems from Edge to cloud, leveraging heterogeneous, chiplet-based computing architectures that can be tailored to specific use cases. The EU CHASSIS project is enabling software-defined vehicles through the development of new architectures and tools.
  • IRT Nanoelec’s Q-Loop program is addressing the challenge of scaling up quantum computer control electronics with error correction. Model-based systems engineering with SysML V2 was used to model a quantum computer system and test it on an attenuation line. The research is now focusing on expanding the models to the entire chain and integrating the visualization and verification of properties.
  • The objective of the twelve-partner UDD@Orano project is to integrate advanced digital technologies to speed up nuclear fuel cycle plant modernization. CEA-List helped develop eleven prototypes, hardware, and software to respond to a variety of critical needs, from locating radioactive sources, counting neutrons, and measuring gamma extremity doses to detecting actinide solutions, analyzing complex gamma spectra, and recalibrating simulation models.