Course syllabus
Welcome to STA320 Machine learning and AI!
Study period: November 4th 2026 - January 17th 2027
Registration period: October 28th - November 11th 2026
Further information can be found in the course modules.
| Week | Monday — 4 h block | Monday — 2 h block | Wednesday — 4 h block |
| 1 | Learning a Classifier from Data | From Gradient Descent to a Single Neuron | Training Loops and Project 1 Launch |
| 2 | From One Neuron to a Neural Network | Backpropagation and Model Flexibility | Seminar 1: How Neural Networks Learn |
| 3 | From R to Python: Tensors, Models, and Loss Functions | Training Neural Networks in PyTorch | Applied Python Modelling and Project 1 Clinic |
| 4 | Generalization and Diagnosing Overfitting | Regularization and Justified Complexity | Seminar 2: Evidence for Generalization |
| 5 | Convolution and Feature Maps | Building a Compact Convolutional Neural Network | CNN Modelling and Project 2 Launch |
| 6 | CPU-Efficient Image Modelling | Interpretation, Error Analysis, and External Generalization | Seminar 3: Image-Project Progress |
| 7 | Attention and the Transformer Architecture | Project 2 Group Vivas | — |
| 8 | Pretrained Transformers and Modern AI in Biostatistics | Seminar 4: Model Choice, Ethics, Sustainability, and Exam Review | — |
Important documents:
Schedule - The schedule is preliminary and may change up to two weeks before the course starts
Contact information
Course leader and examinator: Martin Adiels, martin.adiels@wlab.gu.se
Course administrator: Anna Westblom, appliedbiostatistics@gu.se
Visiting adress: Guldhedsgatan 5A, plan 4, Göteborg (for visits, please make an appointment by e-mail or telephone)
Postal adress: Göteborgs universitet, Hälsovetenskapligt centrum, Box 453, 405 30 Göteborg
General student information:
Registration, results and certificates