Learning paths

Choose a route that matches your goal. Follow its tutorials in order, or jump to the topic you need.

New to PyKale

Start here. The gentlest path from a first run to a working domain-adaptation model.

1. Domain adaptation on synthetic data

2. Domain adaptation on digit images

3. Efficient video data loading

Transfer & domain adaptation

PyKale’s core strength – aligning models across shifting data distributions.

1. Domain adaptation on digit images

2. Multi-source domain adaptation on Office

3. Domain adaptation for action recognition

4. Multi-site neuroimaging classification

Multimodal learning

Combining modalities – audio and image, graphs and sequences, multiomics.

1. Audio-visual digit classification

2. Drug-target interaction prediction with DrugBAN

3. Multiomics cancer classification with MOGONET

Biomedical AI

Applications across protein, cancer and cardiac data.

1. Drug-target affinity prediction with DeepDTA

2. Multiomics cancer classification with MOGONET

3. Cardiac MRI diagnosis with multilinear PCA

4. Uncertainty estimation for landmark localisation