Speaker
Andre Lukas
(University of Oxford)
Description
I will discuss how to compute Yukawa couplings in realistic heterotic string models, using machine learning techniques. More specifically, I will explains how training of neural networks can be used to find the main geometrical ingredients of string theory Yukawa couplings, namely approximate Ricci-flat metric on Calabi-Yau manifolds, Hermitian Yang-Mill connections and harmonic bundle-valued forms. The techniques are illustrated for a specific string model with a low-energy standard model spectrum.
