Machine Learning Driven Accelerated Composites Design

We integrate computational physics, molecular dynamics, and multi-fidelity deep learning to rapidly predict and optimize material behavior, enabling faster and more accurate composites design. Our work advances damage-tolerant architected ceramics, accelerates property extraction from atomistic simulations using neural networks, and pioneers nanoscale functionally graded materials through first-of-its-kind MD studies. Together, these efforts establish a powerful data-driven framework for designing next-generation composites with tailored performance and unprecedented efficiency. 

Damage-Tolerant Architected Ceramic

Deep Learning in Materials Science 

Functionally Graded Materials