Chairs of Statistics and Econometrics

Johannes Brachem

Statistics PhD student

Research
I work on the development of the Liesel framework for probabilistic programming, a Python framework for research on complex Bayesian modelling. I gratefully acknowledge the funding provided by the German Research Foundation (DFG) for the development of Liesel through grant 443179956.
In my PhD research, I work on Bayesian Conditional Transformation Regression Models, a very flexible class of distributional regression models. The goal of these models is to provide regression models that can capture all moments of the response's distribution and relate them to covariates - all without the assumption of a fixed parametric distribution.

Publications

  • Brachem, J., Wiemann, P.F.V., Katzfuss, M. (2026) Data-Efficient Generative Modeling of Non-Gaussian Global Climate Fields via Scalable Composite Transformations. arXiv. https://doi.org/10.48550/arXiv.2602.23311
  • Herp, M., Brachem, J., Altenbuchinger, M., Kneib, T. (2025) Graphical Transformation Models. arXiv. https://doi.org/10.48550/arXiv.2503.17845
  • Brachem, J., Wiemann, P.F.V., Kneib, T. (2025) Bayesian Penalized Transformation Models: Structured Additive Location-Scale Regression for Arbitrary Conditional Distributions. arXiv. https://doi.org/10.48550/arXiv.2404.07440
  • Brachem, J., Frank, M., Kvetnaya, T., Schramm, L. F. F., & Volz., L. (2022). Replikationskrise, p-hacking und Open Science – Eine Umfrage zu fragwürdigen Forschungspraktiken in studentischen Projekten und Impulse für die Lehre. Psychologische Rundschau, 73(1), 1-17. https://doi.org/10.1026/0033-3042/a000562
  • Brachem, J., Heitzig, J., Keser, C. (2023) Election Risk in Climate Negotiations – a Lab Experiment with a Non-Convex Bargaining Set. Preprint. http://dx.doi.org/10.2139/ssrn.4624846
  • Brachem, J., & Rothe, A. (2021). Stop removing stop words – An evaluation of preprocessing techniques for Twitter sentiment analysis with a deep learning approach. In R.-M. Kruse, B. Säfken, A. Silbersdorff, C. Weisser (Eds.), Learning deep textwork – Perspectives on natural language processing and artificial intelligence (pp. 37 – 53). Universitätsverlag Göttingen. https://doi.org/10.17875/gup2021-1608
  • Brachem, J., Krüdewagen, H., Hagmayer, Y. (2019). The Limits of Nudging: Can Descriptive Social Norms Be Used to Reduce Meat Consumption? It's Probably Not That Easy. PsyArXiv. https://doi.org/10.31234/osf.io/xk58q.


Software


Teaching (current)

  • M.WIWI-QMW.0001: Generalized Regression (Lecture) | Summer Term 2026
  • M.WIWI-QMW.0037: Advanced Bayesian Inference (Exercise class) | Winter Term 2025/26


Teaching (past)

  • M.WIWI-QMW.0037: Advanced Bayesian Inference (Exercise class) | Winter Term 2025/26
  • M.WIWI-QMW.0037: Advanced Bayesian Inference (Exercise class) | Winter Term 2024/25
  • M.WIWI-QMW.0037: Advanced Bayesian Inference (Exercise class) | Winter Term 2023/24
  • B.WIWI-QMW.0001: Lineare Modelle (Exercise class) | Summer Term 2023
  • M.WIWI-QMW.0002: Advanced Statistical Inference - Likelihood and Bayes (Exercise class) | Winter Term 2022/23
  • B.WIWI-QMW.0001: Lineare Modelle (Exercise class) | Summer Term 2022


Funding and awards

  • 2025: Best Student Presentation Award for the talk Scalable composite transformations for generative climate model emulation at the 39th International Workshop on Statistical Modelling in Limerick, Ireland.
  • 2024: Second place at the student poster competition of the IRSA Conference (New Perspectives on the Analysis of Complex Multivariate Data) in Minneapolis, MN, USA, with the poster: Chaining transformation models and Bayesian transport maps for high- dimensional non-gaussian spatial fields.
  • 2024: Six-month research stay with Matthias Katzfuss at the University of Wisconsin–Madison, funded via a scholarship by the DAAD (German Academic Exchange Service)


Topics for Bachelor's and Master's theses

Short description: Structured additive distributional regression models allow flexible, component-wise specification of all parameters of a response distribution, such as mean, scale, and shape, using structured additive predictors based on penalized splines and other smooth components. While full Bayesian inference is attractive for uncertainty quantification, it becomes computationally demanding for large datasets or complex model structures. To address this challenge, this thesis integrates Stochastic Gradient Hamiltonian Monte Carlo,a scalable gradient-based Markov chain Monte Carlo method that leverages minibatch stochastic gradients into the distributional regression setting. Prior programming experience in Python is an essential prerequisite for this project.

Relevant literature:

Master's Thesis

Contact: Johannes Brachem (brachem@uni-goettingen.de)

Short description: Penalized transformation models (PTMs) are a semiparametric location-scale regression family that estimate a response's conditional distribution directly from the data, and model the location and scale through structured additive predictors. The core of the model is a monotonically increasing transformation function that relates the response distribution to a reference distribution. One current limitation for Bayesian PTMs is slow Markov Chain monte carlo (MCMC) sampling, making large datasets challenging for PTMs. This thesis develops a fast and numerically robust Iteratively Re-Weighted Least Squares (IWLS) sampler for PTMs. Prior programming experience in Python is an essential prerequisite for this project.

Relevant literature:

Master's Thesis

Contact: Johannes Brachem (brachem@uni-goettingen.de)

Short description: Penalized transformation models (PTMs) are a semiparametric location-scale regression family that estimate a response's conditional distribution directly from the data, and model the location and scale through structured additive predictors. The core of the model is a monotonically increasing transformation function that relates the response distribution to a reference distribution. This thesis extends Bayesian PTMs to count data. Prior programming experience in Python is an essential prerequisite for this project.

Relevant literature:

Master's Thesis

Contact: Johannes Brachem (brachem@uni-goettingen.de)

Short description: Linear constraints can be incorporated into Bayesian structured additive models by reparameterizing the model, or by directly conducting constrained sampling without changes to the model. This thesis implements constrained sampling in Python in the probabilistic programming framework Liesel and compares the performance of this implementation to constraints via reparameterization in terms of speed, numerical stability, and scalability. Prior programming experience in Python is an essential prerequisite for this project.

Relevant literature:

Master's Thesis

Contact: Johannes Brachem (brachem@uni-goettingen.de)

Short description: This thesis investigates simulation-based calibration (SBC) as a diagnostic tool for Bayesian additive distributional regression models. Additive distributional regression extends classical regression by modeling not only the mean but all parameters of a response distribution, such as scale, shape, or skewness, through structured additive predictors. This flexibility allows rich, data-driven modeling. SBC is an approach based on repeated simulation from the prior and refitting the model to assess whether posterior inferences are well-calibrated. The thesis outlines the SBC procedure, discusses practical choices such as summary statistics and rank-based diagnostics, and adapts it to the specific structure of additive distributional regression. Through simulation experiments, the thesis evaluates how violations of model assumptions, prior choices, or numerical issues affect calibration.

Relevant literature:

Bachelor's or Master's Thesis

Contact: Johannes Brachem (brachem@uni-goettingen.de)