Machine Learning in Materials Science

Graduate / advanced undergrad

Graduate / advanced undergrad

Semester: Fall ｜ Audience: graduate students in chemistry, materials, and related interdisciplinary fields

Course Introduction

This course is designed for researchers in the material sciences. It introduces methods and applications of machine learning and deep learning in molecular representation, property prediction, retrosynthesis, and materials discovery. Balancing algorithmic principles with research practice, it helps students truly apply AI tools to their own work—from understanding a methods paper to running a model on their own system.

Key Topics

• Graph neural networks—learning structure–property mappings from molecular graphs.
• Property prediction—second-scale prediction from spectra to physicochemical properties.
• Retrosynthesis—data-driven reasoning of feasible synthetic routes.
• Materials discovery—property-constrained generation and optimization of molecules and materials.

Teaching Format

Lectures + code practice + TBD (course projects / literature reports, etc.).

For reference materials or lecture invitations, feel free to write.