Long dormant, Artificial Intelligence (AI) is back on the scene, propelled by technological progress in both software and hardware. This revival will have a major impact on our lives. But the promise of AI is rivalled only by the research challenges that must be overcome.
As a major research and innovation stakeholder, CEA is deeply engaged in AI, experimenting with AI use cases to solve societal challenges in areas like health, energy, and transportation, and by developing Edge and trusted AIs.
Edge—or embedded—AI (eAI) is a new research field that calls for new hardware platforms and new algorithms.
Algorithms for eAI must take into account a constrained hardware environment and adapt to evanescent data.
Secondly, if AI is to be used in decision-making processes, trust is a key issue. Conventional software engineering practices, such formal methods and model-driven software engineering, must therefore be revisited in the context of AI.
CEA is addressing these challenges through research programs in embedded and trusted AI.