Embedding
AI
processing in intelligent systems and devices
A single community united around the goal of developing an embedded artificial intelligence model that combines sovereignty, trust and frugality.
The DeepGreen project is engaged in developing the Eclipse AIDGE open source software platform
18
Members
1
Software platform
7
Demonstrators
4
Years
Why this project?
More about the platformEmbedding AI processing capability in intelligent systems and devices is a megatrend that signals a major transformation of the digital world.
Embedded AI offers solutions for processing data as closely as possible to the data source, which has a number of innate advantages: faster processing, reduced data transmission, and higher levels of data confidentiality and security.
These sovereign technologies are essential if we are to retain our independence, compete at a higher level and ensure that our solutions comply fully with our standards and regulations.
Our ambition: to make embedded AI a major strength of France and the wider Europe.
More about the platform
The project members
A group of around twenty major French partners from research and industry committed to pooling their skills for the express purpose of designing the Eclipse AIDGE platform.
- Adagos
- Airbus
- ArcelorMittal
- Arcys
- CEA
- CS Group
- Dassault Aviation
- Dolphin Design
- EDF
- Ezako
- Hawai.Tech
- Inria
- Kalray
- MBDA
- Onera
- Pulse audition
- Sysnav
- Thales
Our challenges
The goal of DeepGreen is to bring together a group of national stakeholders to create and share a software platform dedicated to the design and large-scale deployment of embedded AI applications offering innovative and value-creating functions in response to the following strategic challenges:
Sovereignty
Providing control over the end-to-end deployment of embedded AI and promoting French and European hardware technologies.
Competitiveness
Accelerating the rollout of embedded AI applications by giving all French and European R&D stakeholders access to the best-possible development tools.
Frugality
Developing advanced model optimisation techniques to reduce memory footprint and energy consumption.
Reliability
Developing robust and explainable approaches alongside methods for evaluating and qualifying the critical functions required to deliver the performance expected of embedded AI.
An ecosystem of partners
More about the ecosystemA project supported by
Follow all the project news