aller au menu | aller au contenu
Objectives : Manufacturers developing critical systems are in a hurry to take advantage of the performance of modern AI to pave the way for the introduction of new capabilities into their systems. But performance alone is not enough when it comes to critical systems with qualification or even certification constraints. AI must then have properties such as: robustness, guarantees, explainability, etc. They must "open the box" and understand its contents, qualify their systems, anticipate the creation of standards or certification requirements.
This training aims to make participants aware of the limitations and possibilities of AI for critical systems.
Public concerned : senior engineers and technicians
Prerequisites : Bac + 5 with knowledge of mathematics
It is desirable to have a good algorithmic culture and basic knowledge of Artificial Intelligence techniques.
Acquired skills during the training : At the end of the training, participants will be able to:
Understand the specific issues of certification or qualification of AI-based systems;
Know the limits of the current state of the art in machine learning;
Know the main normative and research initiatives in progress on the subject.
Targeted careers : R&D engineers or design office
Notes : existing continuing education
maximum number: 12 people