Artificial Intelligence - Module 2

I have been teaching this course since 2026.

This course deepens the theoretical and practical understanding of artificial neural networks and their applications in biomedical domains. It introduces key concepts behind modern deep learning architectures, focusing on how these systems learn, generalise to unseen data, and generate new data. Students will explore most of the existing neural network classes, including convolutional, recurrent, and transformer-based architectures, with an emphasis on biomedical data, including such multi-omics data and images. The course also covers regularisation strategies, weight initialisation, and the effects of overparameterisation, providing insights into the bias-variance trade-off and the double descent phenomenon. In the second part, students will be introduced to unsupervised learning, focusing on deep generative models, including Generative Adversarial Networks (GANs) and Variational Autoencoders (VAEs), and their applications in biomedical research. Practical sessions in Python, using PyTorch and related tools, will enable students to implement, train, and evaluate their models on real-world biomedical datasets.

For more information, visit the course page.