Guided Research
What I do
For application in safety-relevant and sensible fields, AI need to be trustworthy. To achieve that goal, interpretable ML models are required, that report how their prediction is performed and how certain they are. In summersemester 2020 I conducted research on uncertainty estimation of machine learning models. With the inspriring and motivating guidance of Dr. Shadi Albarqouni and Agnieszka Tomczak, I explore the novel field of labels uncertainty within an extensive study of epistemic, aleatoric and Out-Of-Distribution (OOD) uncertainty.
What I learn
- Independent research work
- Work towards a grand vision
- What is my drive for work?