Speaker:
Patrick Vaudrevange
Institution:
TU Munich
Date:
Wednesday, January 13, 2021
Time:
11:00 am
Location:
Zoom Seminar
Abstract:
The landscape of four-dimensional string models is of enormous size and, hence, widely undiscovered. Therefore, it is reasonable to assume that there are many islands of phenomenological promising string models to be found in this landscape. In this talk, an overview in the case of the heterotic orbifold landscape is given using various techniques from machine learning. In more detail, we present i) an autoencoder neural network to identify structures in this landscape, ii) contrast patterns to construct new MSSM-like string models and iii) neural networks to predict the stringy origin of the MSSM. Moreover, by analyzing the string landscape some novel ideas on flavor, CP and dark matter will be uncovered.
The landscape of four-dimensional string models is of enormous size and, hence, widely undiscovered. Therefore, it is reasonable to assume that there are many islands of phenomenological promising string models to be found in this landscape. In this talk, an overview in the case of the heterotic orbifold landscape is given using various techniques from machine learning. In more detail, we present i) an autoencoder neural network to identify structures in this landscape, ii) contrast patterns to construct new MSSM-like string models and iii) neural networks to predict the stringy origin of the MSSM. Moreover, by analyzing the string landscape some novel ideas on flavor, CP and dark matter will be uncovered.
Host:
Michael Ratz