From atomically resolved imaging to generative and causal models

The development of high-resolution imaging methods such as electron and scanning probe microscopy atomic tomography have provided a wealth information on structure functionalities solids. availability this data in turn necessitates approaches to derive quantitative physical information, much like the scattering early XX century which given one most powerful tools condensed matter physics arsenal. Here, we argue that transition requires adapting classical macroscopic definitions, can enable fundamentally new opportunities understanding chemistry. For example, many definitions symmetry be introduced locally only Bayesian sense, balancing prior knowledge materials' experimental yield posterior probability distributions. At same time, local allows for description solids based construction statistical generative models, akin Ginzburg-Landau thermodynamic models. Finally, note observational opens pathways towards exploring causal mechanisms underpinning solid functionality.