Statistical physics of disordered systems
Author：Yuki Rea Hamano
Affiliation：Department of Physics, Graduate School of Science, The University of Osaka
Abstract：We study wetting transition phenomena in a family of multi-layer spin glass (SG) models using replica theory in the dense limit and perform large-scale replica exchange Monte Carlo (REMC) simulations of the corresponding finite-connectivity models to substantiate the theory.
Theoretically, we observe a rich wetting phenomenology for both the ferromagnetic (FM) and replica symmetry broken (RSB) SG models: stable branches of solutions are identified, some of which do not have a bulk counterpart. Importantly, we consistently find stable spatially heterogeneous RSB phases that are novel to SG models.
The REMC simulations confirm the existence of these heterogeneous phases. Notably, the spatially heterogeneous RSB phases persists even in the finite-connectivity SG models in the thermodynamic limit. The RSB phases found in this microscopic model of SG is reminiscent of those found in a model of neural networks by Yoshino.
Affiliation：Department of Physics, Graduate School of Science, The University of Osaka
Abstract：We study wetting transition phenomena in a family of multi-layer spin glass (SG) models using replica theory in the dense limit and perform large-scale replica exchange Monte Carlo (REMC) simulations of the corresponding finite-connectivity models to substantiate the theory.
Theoretically, we observe a rich wetting phenomenology for both the ferromagnetic (FM) and replica symmetry broken (RSB) SG models: stable branches of solutions are identified, some of which do not have a bulk counterpart. Importantly, we consistently find stable spatially heterogeneous RSB phases that are novel to SG models.
The REMC simulations confirm the existence of these heterogeneous phases. Notably, the spatially heterogeneous RSB phases persists even in the finite-connectivity SG models in the thermodynamic limit. The RSB phases found in this microscopic model of SG is reminiscent of those found in a model of neural networks by Yoshino.
Posted : March 31,2026
