UNC-CH Physics and Astronomy Colloquium

Paulo Bedaque, University of Maryland

“Path integrals, complexified fields and machine learning”

The Feynman path integral is a tool applicable to all fields of physics. Recently, the possibility of changing the “path” of integration to lie on a complex manifold has been explored in both theoretical and numerical contexts. We review recent work using this ideas directly attacking the famous “sign problem” appearing in numerical studies of finite density systems and in the computation of the subtle phase leading to quantum interference. We will provide a short introduction to artificial intelligence techniques that were used in this work.