Research database

The vast majority of modern AI is computationally demanding at all stages, from algorithm design and development, to training and deployment. This is true for all kinds of hardware supporting AI, from cloud and HPC with large scale centralized data storage, to edge devices with limited computational and memory resources. The objective of this Spoke is to lay the foundations for the development of a new generation of sparse, robust, adaptive and accurate AI algorithms running on extreme hardwares, from edge to exascale. These algorithms will support intelligent systems and autonomous agents, with varying degrees of supervision, where applications go from industrial robotics to banking, mobility and defense, energy management and health. The foundational work will be closely coupled to the study of the social, economical and environmental sustainability of the proposed solutions and their applications.