Challenging AI with Challenges from Physics: How to solve fundamental problems in Physics by AI and vice versa

• Unité de recherche: SEFRI
• Numéro de projet: 24.00306

Résumé des résultats (Abstract) (Anglais)

AI is a disruptive technology that is currently changing not only many research fields but also substantially challenges current business concepts. Keeping up with the ever-faster progress on one hand side and the ever increasing competition from US and Chinese IT giants is a challenge for European Industry and Research Institutions. Yet, new concepts are required to keep pace. This concept is to intertwine modern AI concepts with strategies from the application field, a strategy where we claim that we can contribute substantially to a novel and promising research field. Particle physics is a formidable basis for such developments as there are no critical, personal data, data can be easily generated; and we have a good understanding of the underlying models and the ground truth. Yet, particle physics has challenges such as detailed questions of new physics and extremely high precision of the results. We identified three main fields where this project will contribute by solving 1) highly ill-posed inverse problems using physics models, 2) handling uncertainties, rare events and give reliable error bounds, and finally, 3) to be able to give explanations of the machine learning results in terms of physics ontologies and models. This highly challenging research agenda is tackled by 9 internationally recognized and leading researchers from both physics and computer science, from 5 universities in 5 European countries, all being member of the 4EU+ Alliance of European Universities. Over 4EU+ an excellent infrastructure for unique training opportunities is available to enable the research fellows to reach the ambitious goals, on one hand side and to educate them in critical and innovative thinking, management and social skills to prepare them optimally for leading positions in academics and industry.