The vision of FedEU.ai – Connecting AI Knowledge – is a federated cloud-edge AIcontinuum based on privacy-preserving and federated computing elements. In such a cloud environment, private data, data processing (AI) services and insight generation can be fully decoupled from each other. Besides, algorithms can access data or machine learning algorithms can train on encrypted data. Privacy-preserving data spaces can be flexibly deployed, whether data resides on cloud, edge or on-premises, and the environmental impact of AI can be reduced.
The project aims to create a facility for exploring novel technologies that can facilitate federated cloud AI, the translation of research to large scale industrial exploitation, and to develop an industry-scale federated AI technology stack fully harmonised with EU cloud initiatives and existing solutions. FedEU.ai will enable the agile deployment of data services on distributed data, using the combination of state-of-the-art federated learning and privacy-preserving technologies, thus greatly increasing the output of European data economy without exposing EU data assets.
The FedEU.ai solution extends the potential for innovative data (re)use and services by cost-effectively expanding data resource utilisation and analysis based on an independent cloud platform to enable any type of cloud deployment. FedEU.ai will substantially reduce the environmental impact of large-scale machine learning. The FedEU.ai innovation consists of:
- scalable, cloud-edge continuum federated AI services,
- decoupling data from processing, and insight generation on the cloud-edge,
- energy efficient cloud-edge AI, reducing the environmental impact of AI models,
- advanced federated security & privacy protocols, as well as
- trustworthy data and models throughout the federated learning process.