Distributed AI in Supercomputing and phYsics

The Distributed AI in Supercomputing and phYsics (DAISY) project is a joint undertaking of the Departments of Computer Science and Physics at the University of Helsinki. It is led by Finland’s top experts in artificial intelligence (AI), machine learning, space plasma physics, and supercomputer simulations. DAISY proposes to develop groundbreaking AI algorithms utilising graphical models as a basis for topological and symbolic data representations. These algorithms can detect signatures of the fundamental physical processes at the heart of the dynamics of solar eruptions, auroras, and fusion power reactors. The goals are scalable, explainable, and reliable AI methods that can be distributed in supercomputers in resource-efficient ways. The new methods, developed first within plasma physics, will revolutionise the fields of AI and high-performance physical simulations at large. DAISY represents an important vanguard case combining high-performance computing with artificial intelligence.