Syllabus

Course overview¶

This course introduces practical AI tools for materials science, blending lectures with hands-on notebooks. Students will learn to frame scientific questions as machine learning and data science tasks, build reliable models, and integrate them with simulation or experimental pipelines.

Learning objectives¶

• Formulate materials problems for supervised, and unsupervised.

• Engineer features for molecules, crystals, and microstructures.

• Train, tune, and evaluate models.

• Communicate results with reproducible workflows and documentation.

Schedule¶

• Week 1: Linear models

• Week 2: Kernel models

• Week 3: Neural Networks

• Week 4: Gradient descent and differentiable programming

• Week 5: Molecular representations

• Week 6: Graph Neural networks

• Week 7: Reading week break. No classes this week!!

• Week 8: MACE or PySCF tutorial

Final project¶

Due to the length of the course, the grading will be based on a final project that includes three parts.

1. Written Report, 4 Pages maximum

   1. Page 1: Introduction

   2. Page 2: Methodology

   3. Pages 3-4: Results, Discussion and Summary.

   4. Extra pages include references, and additional data and results.

2. Oral Presentation, 20 min maximum

   1. 20 min long.

   2. Clearly state what is the goal of the project

   3. Methodology description

   4. Some results discussion

   5. Conclusion

3. Code

   1. Jupyter notebook file (.ipynb) deployable in Google Colab.

   2. It should run smoothly and self-explanatory for other students.