Supervisors: Christoforos Rekatsinas, Vasilis Sioros, Panagiotis Krokidas
Description:
Metal-Organic Frameworks (MOFs), and with architected materials, hold significant promise for advanced industrial applications, including gas separation, H2 adsorption, enhanced ductility, and damping. Despite progress in predictive methods, challenges persist in designing both MOFs and architected composites with superior functional properties, such as high adsorption capacity, mechanical resilience, and energy dissipation. While deep learning (DL)-based generative models have shown potential for molecular and material design, their reliance on large datasets limits their applicability in data-scarce domains. Grammar-based generative approaches offer an interpretable and data-efficient alternative, enabling the explicit integration of chemical and structural constraints. This project aims to adapt a grammar-based generative method for the design of MOFs and architected materials, incorporating property prediction models to optimize structures for targeted applications, including gas separation, hydrogen storage, mechanical durability, and damping performance.
Qualifications required: Python programming, Machine Learning and Deep Learning algorithms
References:
crek[at] iit [dot] demokritos [dot] gr