I finished my master in mathematics at the TU Berlin and gained different experiences in applied mathematics and software engineering: Please checkout the attached image if you want to know more about my background!
Although the connection of mathematics and medicine might not be obvious on the first sight, already the ‘father of modern medicine’ William Olster knew:
‘Medicine is a science of uncertainty and an art of probability.’
In my PhD I am going to apply probabilistic deep learning methods to histopathology within the CLARIFY project. Although deep learning has shown impressive results in medical image classification, there remain two major sources of uncertainty: Firstly, the annotation of the datasets as pathologists have a different grade of experience and even expert pathologists can disagree in the disgnosis. Secondly, the deep learning model itself can introduce errors due to limitations of the model architecture, dataset size etc. We want to handle this uncertainty by using crowdsourcing methods and probabilistic deep learning models that are able to estimate this uncertainty. This is important in clinical practice to know when doctors can trust the AI diagnosis -and when not. And what are the basics of these probabilistic methods? Correct, mathematics!
I am excited what we can achieve in this direction together with my supervisor Prof. Rafael Molina and the other CLARIFY researchers.
Arne discovering the Sierra Nevada, Granada (Spain)
Arne Schmidt – ESR8