The rapid advancement of computational materials science and the emergence of deep learning technologies are transforming the landscape of materials discovery, design, and optimization. This session focuses on the integration of physics-based simulations, high-throughput computations, and data-driven approaches to accelerate innovation in materials research. A particular emphasis will be placed on modeling of materials microstructures, including their generation, evolution, and optimization, as well as the theoretical modeling of diverse functional materials beyond structural mechanics. These include functionally graded materials (FGMs), thermoelectric materials, metal-organic frameworks (MOFs), and piezoelectric materials, where the interplay between structure and electronic functionality is critical. The related crystal structures modeling and phonon engineering are also depicted. Topics of interest include, but are not limited to: • Theoretical modeling and design of functional materials such as functionally graded materials (FGM), thermoelectric materials, and metal-organic frameworks (MOFs). • Microstructure modeling and optimization using mathematical models and AI-assisted techniques • Applications of deep learning in microstructure-property relationships, defect prediction, and synthesis planning • Autonomous materials discovery using active learning and reinforcement learning In this session, through interdisciplinary dialogue, we aim to define the next-generation paradigm of materials research empowered by advanced computation and deep learning methodologies.