Adding loops to mean field approximation for disordered models
Federico Ricci-Tersenghi, Université Sapienza, Rome
In disordered models, the extension of mean-field approximations in order to include the effects of short loops is a challenging task, still largely open. Recently we have developed an analytical approximation (Replica Cluster Variational Method) and some related algorithms, which are able to provide a much better description of finite-dimensional spin glass models, than what the Bethe approximation does. I will summarize these results.