Dr. Andreas Lintermann, Invited Lecture, 70th Anniversary CERN, CERN, Geneve, Switzerland:
With the continuously increase in availability of data, e.g., produced by the collision experiments at CERN already nowadays and the extreme multiplication of the data volume CERN is facing with their future experiments, more mature technologies are needed to find the needle in the haystack and to make new physics discoveries. Obviously, these amounts of data cannot be analyzed anymore with traditional methods. Artificial Intelligence (AI) technologies more and more penetrate all different fields of science and the immense data availability together with heterogeneous High-Performance Computing (HPC) systems open the gate for efficient AI-based tools to find unseen correlations, detect the unknown, accelerate computation, or model otherwise complex processes. This talk discusses the future challenges of scientific communities and that AI technologies need to develop towards foundational, multi-purpose models applicable to a wide set of applications. It also discusses the costs of training such AI, how HPC can help in speeding up the computations, and what funding instruments and infrastructure the European Union is providing to academia and industry to cope with such challenges.
