# Paul Ungermann

> PhD student in AI4Science at the Technical University of Munich (TUM), supervised by Prof. Nils Thuerey in the Physics-based Simulation group. Works on foundation models for partial differential equations (PDEs): pretraining one neural surrogate on simulations from many physical systems so it can be adapted to a new problem from a few examples. Also interested in uncertainty quantification and reliable machine learning.

Website: https://paulungermann.github.io/
Email: paul.ungermann@tum.de
Based in Munich, Germany.

## Profiles

- [Google Scholar](https://scholar.google.de/citations?user=-dmbOmQAAAAJ)
- [GitHub](https://github.com/paulungermann)
- [LinkedIn](https://www.linkedin.com/in/paul-ungermann)
- [X](https://x.com/PaulUngermann)

## Publications

- [Diffusion LLMs are Natural Adversaries for any LLM](https://arxiv.org/abs/2511.00203): Paul Ungermann\*, David Lüdke\*, Tom Wollschläger\*, Stephan Günnemann, Leo Schwinn. NeurIPS 2026. \*Equal contribution. Grew out of his master's thesis.
- [Using machine learning to automate the collection, transcription, and analysis of verbal-report data](https://doi.org/10.3758/s13428-025-02800-5): Tehilla Mechera-Ostrovsky, Paul Ungermann, Chris Donkin. Behavior Research Methods 57(10), 2025. [Code](https://github.com/tehillamo/AutoV-LLM)
- [Robust 3D Gaussian Splatting for Novel View Synthesis in Presence of Distractors](https://arxiv.org/abs/2408.11697): Paul Ungermann, Armin Ettenhofer, Matthias Nießner, Barbara Roessle. GCPR 2024. [Project page](https://paulungermann.github.io/Robust3DGaussians), [Code](https://github.com/paulungermann/Robust3DGaussians)

## Experience

- Nov 2026 to now: PhD student, AI4Science, Technical University of Munich (Prof. Nils Thuerey)
- Jan to Sep 2026: Wireless Data Science Intern, Apple, Munich. Agentic workflows and predictive ML models for cellular performance analytics in the Wireless Analytics and Intelligence team.
- Sep 2023 to Sep 2025: Research Assistant, LMU Munich, Computational Psychology chair of Prof. Chris Donkin. ML pipeline for recording and analyzing verbal reports.
- Aug 2019 to Aug 2021: Web Developer (working student), Wirth & Horn Informationssysteme, Munich

## Education

- Nov 2026 to now: PhD, AI4Science, Technical University of Munich
- Apr 2023 to Apr 2026: M.Sc. Computer Science, Technical University of Munich. Grade 1.1, top 5% of class. Thesis: Guiding Diffusion Models to Attack Large Language Models.
- Oct 2019 to Mar 2023: B.Sc. Computer Science, Technical University of Munich. Minor in mathematics, tandem exchange with Tsinghua Shenzhen International Graduate School. Thesis: Gaussian Processes for Light Microscopy Image Segmentation.

## Talks

- Sep 2024: Robust 3D Gaussian Splatting, GCPR 2024, Munich
- Jul 2024: "Using LLMs to Automate the Analysis of Verbal Reports", MathPsych 2024, Tilburg, Netherlands
- Mar 2024: "Using LLMs to Automate the Analysis of Verbal Reports", TeaP 2024, Regensburg, Germany