Rieck Lab

Our world is full of phenomena that occur at multiple scales. Biomedical research, for instance, commonly observes complex systems at different resolutions, ranging from the macroscopic to the microscopic. Zooming in provides us with the 'fine print' (e.g. individual neurons in a brain), while zooming out lets us see the 'big picture' (e.g. locally-connected networks of neurons, or areas in the brain). For many applications, there is not just one specific scale to consider—relevant features might occur on multiple scales and a priori information about their suitability for a specific task is typically lacking.

With noise being an inevitable part of such investigations, we need tools that enable robust multi-scale analyses. Our research agenda is to create, cultivate, and critique such tools based on topological machine learning techniques, with a specific focus on healthcare topics.

Aims

Develop novel machine learning algorithms based on topological concepts that are (I) aware of the multi-scale nature of complex biomedical data sets, (II) robust to noise, and (III) interpretable.