The SfS-PhD Presentations

The goal of the SfS-PhD presentations is to facilitate the exchange of ideas in an informal atmosphere. Although priority is given to PhD students from the seminar for statistics, we had presentations by various external people. This includes PostDocs, PhD students from other groups and universities, consultants, as well as masters students.

If you are interested in listening in or giving a presentation yourself, please contact me!

Previous presentations

To give you an idea of the topics, here's what happened in the previous three semesters.

Autumn 2023

• Christoph Schultheiss: "On (causal) well-specification of regression estimates".
• Maybritt Schillinger: "Machine learning with unpaired samples – an example of high-dimensional climate projections".
• Felix Kuchelmeister: "Always reject H_0 when the data are separable?".
• Jakob Heiss: "How to forecast consistently based on noisy incomplete observations at irregular observation times".
• Florian Krach: "Path-Dependent Neural Jump Ordinary Differential Equations".
• Simon Briend (Pompeu Fabra & Paris Saclay): "Estimating Adam's Genealogy in Growing Recursive Trees".
• Cyrill Scheidegger: "Fitting high-dimensional additive models under hidden confounding".
• Mathieu Chevalley: "Causal Discovery in the Wild".
• Xinwei Shen: "Distribution is all you need? From generative AI to distributional regression".
• Sorawit Saengkyongam: "Identifying Representations for Intervention Extrapolation".

Spring 2023

• Philipp Schmocker (NTU): "Global universal approximation with functional input neural networks on weighted spaces".
• Qikun Xiang (NTU): "Numerical methods for optimal transport and matching for teams".
• Jakob Heimer: "Thoughts on Visualization (in Statistical Consulting)".
• Felix Schur: "Meta Learning for Lifelong Sequential Decision Making".
• Florian Schwarb: "R-ICP: Invariant Causal Prediction under Random Effects".
• Jakob Heiss: "Bayesian Optimization-based Combinatorial Assignment".
• Sorawit Saengkyongam: "Effect-invariant mechanism for policy generalization".
• Felix Kuchelmeister: "Logistic Regression with small noise or few samples".
• Harald Besdziek: "Nearly parametric rate of the maximum likelihood estimator in mixtures of power series distributions".
• Maybritt Schillinger: "Simulating Regional and Global Climate Models with Machine Learning".
• Juan Gamella: "The Causal Room: datasets for causal discovery from well-understood physical mechanisms".
• Lea Tamberg: "Prediction of wellbeing outcomes based on indicators for basic human needs satisfaction".

Autumn 2022

• Malte Londschien: "On the Generalized Method of Moments".
• Jeffrey Näf: "R-NL: high-dimensional covariance estimation when tails are heavy".
• Christoph Schultheiss: "Ancestor regression in linear structural equation models".
• Juan Gamella: "Validating Causal Discovery Algorithms: Realistic synthetic data with known ground truth".
• Jinzhou Li: "Root cause discovery in linear structural equation models with unknown causal ordering".
• Maybritt Schillinger: "Distribution Shifts in Climate Observations - Just a Mean Change?".
• Felix Kuchelmeister: "Linear Separability: From Dandelions to Logistic Regression".
• Jakob Heiss: "How Infinitely Wide Neural Networks Benefit from Multi-task Learning - an Exact Macroscopic Characterization".