Joren Brunekreef
I am a postdoctoral researcher at the Netherlands Cancer Institute (NKI) and the University of Amsterdam (UvA), where I am part of Prof. Jan-Jakob Sonke’s radiotherapy department and Dr. Jonas Teuwen’s AI for Oncology lab.
My current research interests include quantifying uncertainty in deep learning and, more concretely, improving the calibration of image segmentation models to enhance their reliability in a medical setting. I approach these challenges through the lens of conformal prediction, a framework offering statistically valid prediction intervals in a model-agnostic manner.
I recently completed a PhD in nonperturbative quantum gravity under the guidance of Prof. Renate Loll at the Radboud University Nijmegen. My work primarily involved investigating curvature observables in the context of Causal Dynamical Triangulations (CDT). During this period, I co-developed open-source codebases for performing Markov chain Monte Carlo simulations of CDT in two and three dimensions.
Additionally, I briefly worked as a junior researcher in the infectious diseases modeling group at the UMC Utrecht’s Julius Center. Collaborating with Dr. Alexandra Teslya and Prof. Mirjam Kretzschmar, I helped develop an agent-based network model for simulating societal lockdowns in response to infectious disease outbreaks.
selected publications
- JMRIAI applications to breast MRI: today and tomorrowJournal of Magnetic Resonance Imaging, Apr 2024
- CVPRKandinsky Conformal Prediction: Efficient Calibration of Image Segmentation AlgorithmsIn Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, Apr 2024
- medRxivImpact of adherence and stringency on the effectiveness of lockdown measures: a modelling studymedRxiv, Jan 2024