Disentangled representation learning has been proposed as an approach to learning general representations. This can be done in the absence of annotations, or with limited annotation. A good general representation can be readily fine-tuned for new target tasks using modest amounts of data. This alleviation of the data and annotation requirements offers tantalising prospects for tractable and affordable healthcare. Finally, disentangled representations can offer model explainability, increasing their suitability for real-world deployment.
In this half-day tutorial, a satellite event in conjunction with MICCAI 2020, we will offer an overview of representation learning and disentangled representation learning and criteria, and discuss applications in medical imaging. We will conclude with open ended challenges.