String data and machine learningAndre Lukas, University of Oxford, 12:00 EDT
Abstract: Recently, string theorists have started to use machine learning techniques. I will explain why string theory might be an interesting arena for modern computational methods and discuss two specific applications of machine learning to problems which arise in the context of string model building. In the first part, I will show how machine learning can help understand certain mathematical structures, specifically the cohomology of line bundles on complex manifolds. In the second part, machine learning is applied to a data set of string standard models.
During Andre’s talk, there was a question regarding availability of the heterotic line bundle standard model datasets. You can find them here.