Physics (Internal)

Machine Learning and the String Landscape

This event is part of the HET Seminar Series.

I will review efforts to study string theory using techniques from data science, including supervised machine learning and deep reinforcement learning. Most of the talk will focus on studying a particular ensemble of 10^{755} string geometries, which exhibits some universal physical features, has indirect implications for the de Sitter swampland conjecture, and raises a new dark matter concern.