Nico Disch

Nico Disch is a PhD student at the German Cancer Research Center (DKFZ) working on medical sparse and irregular image time series. Previously he did his Bachelor's and Master's at the University of Heidelberg in Physics. His master thesis was about the exact representation of continuous piece-wise linear functions using Deep Neural Networks, in order to solve differential equations on (non) convex domains. Currently he is developing a model that can deal with irregular and sparse medical image time series, e.g. MRI sequences of the same patient over time, in order to predict future images for disease progression.

Projects:

• Anomaly detection in sparse image time series

Interests:

• Transformers
• Time series analysis
• Neural Processes