New methods for old problems: vacua of maximal D = 7 supergravities

TL;DR

This paper introduces a novel method for finding vacua in maximal D=7 supergravity theories by varying the embedding tensor and using AI-inspired optimization techniques, leading to new Minkowski vacua and detailed spectra.

Contribution

It presents a new approach to identify vacua in supergravity by varying the embedding tensor and applying Evolution Strategies, resulting in the discovery of new vacua and analytical mass spectra.

Findings

Discovered two new Minkowski vacua.

Developed a method to reconstruct analytical spectra from numerical data.

Applied AI optimization techniques to supergravity vacuum search.

Abstract

Finding vacua of supergravity theories is an outstanding problem which has been tackled in several ways, and with this work we add a new method to the puzzle. We analyse the scalar sector of maximal gauged supergravity theories in seven space-time dimensions. We look for vacua of the theory by varying the embedding tensor, instead of directly minimising the scalar potential. The set of quadratic constraints arising from this procedure has been solved by means of Evolution Strategies optimisation techniques, also adopted in Artificial Intelligence studies. We develop some methods to reconstruct and obtain analytical results starting from numerical outcomes, thus obtaining the complete mass spectra. In addition to some of the known vacua, we also obtain two new Minkowski vacua.