Abstract:A large class of 4d N=1 supersymmetric gauge theories that are worldvolume theories of D3-branes probing toric Calabi-Yau 3-folds exhibit toric phases that are IR equivalent under Seiberg duality. These theories are realized in terms of a Type IIB brane configuration characterized by a bipartite graph on a 2-torus known as a brane tiling or dimer model. This graph originates from the coamoeba projection of the mirror curve associated to the toric Calabi-Yau 3-fold. When the complex structure moduli of the mirror Calabi-Yau 3-fold are varied, the coamoeba and corresponding brane tilings change their shape, giving rise to different toric phases related by Seiberg duality. In this work, we show how explainable machine learning techniques can be introduced in order to systematically identify the relationship between choices of complex structure moduli and specific toric phases of 4d N=1 supersymmetric gauge theories. We show that machine learning techniques enable us to explicitly map the phase space of 4d N=1 supersymmetric gauge theories corresponding to a given toric Calabi-Yau 3-fold.

Zoom: https://us06web.zoom.us/j/83257513046                                                                                                                                                             Meeting ID: 832 5751 3046   Passcode: EHCGP2024