PhD
Chair of Applied Statistics in Social Sciences, Economics and Business at LMU
Institut für Statistik
Ludwig-Maximilians-Universität München
Ludwigstr. 33
80539 München
Biosketch
Cornelia holds a Bachelor’s degree in Statistics and a Master’s degree in Data Science from LMU Munich. She then started her PhD at the Chair of Applied Statistics in Social Sciences, Economics and Business at LMU to work on the topic of Uncertainty Quantification in Supervised Machine Learning. Since October 2022, Cornelia is a relAI member.
relAI Research
Learning from Uncertainty: A Statistical Approach
Uncertainty is inherent in empirical data and arises throughout the machine learning pipeline, from data collection and measurement to model specification, estimation, and deployment. This thesis develops a statistical framework for understanding uncertainty not merely as noise, but as an informative signal that can reveal structure in data generating processes. It first formalises aleatoric and epistemic uncertainty and shows that linking types of uncertainty to sources is often nontrivial. It then studies human label variation as a form of aleatoric uncertainty, proposing Bayesian methods to infer latent class structure from multiple annotations and discussing implications for active learning. Finally, it extends this perspective to climate science by linking climate variability to aleatoric uncertainty and using generalized additive models to characterize its spatiotemporal structure. Overall, the thesis shows how learning from uncertainty can deepen understanding across domains
Publications
https://arxiv.org/abs/2305.16703
https://aclanthology.org/2024.unimplicit-1.2/
https://aclanthology.org/2025.nlperspectives-1.7/
https://arxiv.org/abs/2604.15067