FASTER AI:Fullständigt & Autonomt Förverkligande av Säkerhets- och Tidskritisk Inbäddad Artif. Intelligens

Purpose and goal:

FASTER AI addresses emergent needs to embed machine learning (ML) inference capabilities within hardware infrastructure of critical importance and use. We focus on hardware utilized widely in telecommunications as well as airborne systems and other vehicles. Expected results and effects:

This project will contribute to the creation of a software methodology for ML on special critical hardware. The software will include a "neural architecture search" library as well as a compiler toolchain that can combine regular programs with ML-inference. Finally, we will demonstrate our methodology and software on two critical-purpose use-cases.

Approach and implementation: Our planned steps include the following steps: -establish project agreements (wp5) -develop software-agnostic libraries (wp1, wp2) -adapt libraries to the use-case hardware at runtime (wp3) -implement demonstrator projects (WP4)