Chemistry and AI

Since 2024, I have taught the Chemistry and AI Master’s course at Radboud University Nijmegen. The lectures for the 2025 edition have just wrapped up (you can find the course details here: 2025).

Chemistry and AI is targeted at students of chemistry and the molecular sciences, and aims to provide students with concepts and practical skills in machine learning, coding and informatics to solve problems in chemistry. The course assumes no foreknowledge besides chemistry. It provides introductory material on standard machine learning and statistical concepts, and Python programming (no prior experience required). As such, this is effectively a crash-course in machine learning taught over seven lectures and practical sessions (around 12 contact hours). The emphasis is on building intuition on key machine learning concepts with direct reference to illustrative chemistry examples such as C–H bond activation, Metal-Organic Frameworks, and solubility prediction. Significant attention is paid to the practical implementations of ML methodology via Python code. A side effect is that students will be introduced to and will practice programming and working with data.

Students who take this course will gain a good foundation in ML which will provide a basis for learning more advanced topics, and will aid their communication with AI/ML/Data Science specialists.

An outline for the 2025 course is given below.

Week 1

• Course introduction
• Molecular representation practical (Python, The RDKit)

Week 2

• Introduction to Machine Learning
• Practical: Numpy, Pandas, MatPlotLib

Week 3

• Regression
• Regression practical

Week 4

• Classification
• Classification practical

Week 5

• Unsupervised methods
• Unsupervised methods practical

Week 6

• Introduction to Neural Networks
• Neural networks practical

Week 7

• Synthetic Planning
• Retrieval Augmented Generation practical (chat with a paper)