Title:Proliferation and prevention of self-loops in ensembles of interacting binary elements

Abstract:Models based on spins or hysterons with appropriately chosen interactions can capture advanced memory effects in complex materials, such as transients in repeatedly compressed crumpled sheets or sequential computing in driven metamaterials. However, unphysical self-loops dominate the response when interactions are chosen randomly, undermining statistical approaches. Here, we uncover the origin of self-loop proliferation in randomly coupled models. We introduce the weakly asymmetric ensemble to suppress self-loops and then develop interaction ensembles that strictly eliminate them. Finally, we use these ensembles to explore the statistics of large systems. Our work highlights the subtle role of interaction symmetries and paves the way for statistical studies of the sequential response and memory effects in complex, multistable materials.