Materials Intelligence & Discovery

Previous Next

The group focuses on:

Generative AI for atomic and molecular design

We apply generative models—including diffusion models—to explore and generate structures like amorphous carbon, nanoporous catalysts, and grain boundaries.

Data-driven spectroscopy interpretation

We use neural networks to connect experimental spectra (e.g., XANES) with structural features in disordered materials.

Structure–property learning

Our models identify how molecular structure and composition affect key performance indicators like catalytic activity, hydrogen storage capacity, or degradability.

Environmentally relevant materials

We predict degradation pathways, bioactivity, and defluorination potential for persistent chemicals like PFAS.

Application-driven insights

Our research supports design goals in clean energy, advanced coatings, separations, and catalysis by linking atomistic modeling with real-world targets.