The field of indoor propagation modeling and scene understanding is witnessing significant advancements, driven by the development of innovative frameworks and techniques. A key direction in this field is the use of physics-informed and sensing-driven approaches to model indoor radio propagation and understand indoor scenes. These approaches aim to overcome the limitations of traditional methods, which often rely on labor-intensive manual modeling and suffer from limited scalability and efficiency. Noteworthy papers in this area include SenseRay-3D, which presents a generalizable and physics-informed end-to-end framework for indoor propagation modeling, and RISE, which introduces a benchmark and system for single-static-radar indoor scene understanding. Wave-Former is also a notable contribution, as it leverages millimeter-wave wireless signals to achieve high-accuracy 3D shape reconstruction for completely occluded objects. Additionally, the use of Bayesian online learning and uncertainty-aware sampling is being explored for efficient RF passive components modeling.