The modeling and simulation of solid materials across multiple scales remains a central challenge in computational science and engineering. From quantum-scale phenomena to macroscopic mechanical behavior, understanding the interplay between different physical scales requires not only sophisticated numerical methods but also rigorous mathematical frameworks. This minisymposium focuses on the integration of multiscale modeling techniques and machine learning approaches for solid-state systems. We aim to explore how recent advances in machine-learned interatomic potentials, surrogate models, and physics-informed learning frameworks are accelerating the development of multiscale methods, enabling accurate and efficient modeling of complex materials with defects, interfaces, or evolving microstructures. Particular emphasis is placed on bridging scales—from atomistic to continuum—through approaches such as coarse-graining, atomistic-to-continuum coupling, and hybrid physics–machine-learning schemes. Contributions that highlight theoretical developments, algorithmic strategies, or application-driven insights are equally welcome. By bringing together applied mathematicians, computational mechanicians, and materials scientists, this symposium fosters interdisciplinary dialogue aimed at advancing both the mathematical foundations and practical capabilities of multiscale modeling. We especially encourage submissions that demonstrate how machine learning can enhance traditional numerical models, improve data efficiency, or quantify uncertainty in multiscale simulations.