Causal fermion systems, an optimization problem and machine learning techniques

by Dr Robert Jonsson (Stockholm University)

Wednesday 2 Nov 2022, 16:00 → 17:00 Europe/Stockholm

Albano 3: 5230 - Xenon (12 seats) (Albano Building 3)

This talk touches on a maybe unusual combination of topics that, I believe, are of interest to many of us. 

• Causal fermion systems (CFS) are a theoretical framework for a unified description of spacetime and its matter content. The approach is mathematically very far developed and rigorous. Hopefully, anyone who ever cared for quantum gravity (even if only secretly) will find the provided overview and introduction to CFS interesting.
• At the heart of CFS lies a variational principle over measures on operator spaces. In its most fundamental, toy model setting it is nicely embedded and visualized on the Bloch sphere. Thus, everyone who has ever worked with spins or likes geometric intuition hopefully will find this part appealing.
• As optimization problems often are, also this problem is more easily stated than solved. Hence, in [arXiv:2201.06382] we applied machine learning techniques (specifically, automatic differentiation and gradient optimization) to advance the numerical study of CFS. I will present the approach we chose and hope to get input and ideas in the discussion with you on how to improve it.

To this end, I aim for a talk informal in style, in the sense of being interactive and accessible to all. Looking forward to seeing you there!