Conditional diffusion model for analyzing quantum fingerprints in conductance fluctuations

Abstract A conditional diffusion model has been developed to analyze intricate conductance fluctuations called universal conductance fluctuations or quantum fingerprints appearing in quantum transport phenomena. The model reconstructs impurity configurations and quantum-interference patterns in the local density of states (LDOS) of mesoscopic nanowire systems from simulated magnetoconductance data. The visualized attention maps show head-dependent spatial patterns for a device configuration with two impurities, while projections of the learned score function illustrate local denoising behavior around a reference LDOS image. These analyses complement the quantitative performance evaluation by providing qualitative insight into aspects of the reconstruction process.