Generating new physics models from machine learning

Abstract

Machine learning applications are dramatically changing many areas of society and science. After a short survey of key machine learning techniques, Sven Krippendorf will discuss some ways on how machine learning can impact fundamental theoretical physics. One example which we initiated recently (arXiv:1809.02612), looks at the automatisation of constructing physical models, satisfying both experimental and theoretical constraints.

He will present a framework which allows the generation of effective field theories using generative adversarial networks. We identify consistent examples generated by the machine which fall outside the class of data used for training. As a starting point, we apply this idea to the generation of supersymmetric field theories. In this case, the machine knows consistent examples of supersymmetric field theories with a single field and generates new examples of such theories. In the generated potentials we find distinct properties, e.g. the number of minima in the scalar potential, with values not found in the training data. He will comment on further applications of this framework.