Machine Learning for Materials Design

The graduate-level introductory course on "Machine Learning for Materials Design" will explore the principles, methodologies, and applications of machine learning (ML) in the domain of materials science and engineering. By bridging the gap between traditional materials design approaches and the cutting-edge field of machine learning, we will learn the tools and knowledge necessary for data-driven discovery and development of materials.

The following ML aspects are first covered through lectures and live coding:

- Introduction to PyTorch

- Automatic differentiation

- Gradient-based optimization

- Best practices in machine learning

- Feed-forward neural networks

- Convolutional neural networks

- Recurrent neural networks

- Autoencoding neural networks

- Inductive biases

Next, a wide range of materials science and engineering applications of the above ML topics will be explored via hands-on and latest research-oriented group projects. Some representative examples include (may vary during the course):

- Predicting mechanical behavior of composites and polymers

- Identifying molecules with unique properties

- Designing metamaterials with tailored mechanical behavior

- Characterizing imperfections and defects in additive manufacturing through data and physics