EGIA - EdGe technologies for Industrial distributed AI applications

Objective:

EGIA aims to research and develop the technologies necessary to make it practically feasible to run industrial AI applications on the Edge, i.e. on (or near) the cyber-physical devices that generate the data. This objective requires (1) specific AI techniques that take into account the distributed and resource-constrained nature of the Edge (e.g. Federated Learning is particularly important as the distributed nature of Edge nodes enhances this paradigm), (2) high-performance resilient communications middleware that adapts to a more hostile environment than the cloud, and (3) DevOps methods and techniques specifically oriented to this type of applications (MLOps, AIOps). All this to enable the creation, optimisation and deployment of distributed AI/ML pipelines with guaranteed quality of service, trust, security and privacy.