Experimental techniques such as high-resolution, multi-contrast MRI, 4D light-sheet microscopy, or single-cell sequencing provide unprecedented, unique insights into life's inner workings across scales. However, these methods also yield enormous amounts of data with complex structure and dependencies that are often impossible to analyse manually. Our research focuses on computational approaches to quantify and visualise those patterns and model the processes underlying form and function of complex biological systems. We are currently interested in self-supervised representation learning using deep neural networks to extract robust, invariant structures from raw data that can be shared and translated across modalities. We would further like to connect these statistical approaches with formal knowledge representation and reasoning. The goal is to advance our conceptual understanding and extend the range of computational tools in scientific and clinical practice.
Available PhD projects
- Self-supervised learning for analysis and classification of neuroscience data.
- Computational mapping of cortical microstructure.
- Interactive visualisation of complex biomedical data.
- Bayesian analysis of spatio-temporal patterns.
- From data to knowledge: Connecting statistical pattern analysis and formal knowledge representation.