Skip to content

Hi! I'm Frederik Dennig, a lecturer at the University of Konstanz and a member of the Data Analysis and Visualization Group.

Highlights

I worked on multiple research projects in the public and private sector. The following overview presents a selection of public projects and prototypes that I created or helped to actualize.

About Me

Experience

Eight years at the intersection of machine learning, data analysis, and visualization — currently as a PostDoc researcher and lecturer with the Data Analysis and Visualization Group at the University of Konstanz. I have managed three international research projects (EU H2020, DFG), secured around 300 kEUR in third-party funding, supervised six B.Sc. and M.Sc. theses, and published 30 peer-reviewed papers. Before the PhD, a software-development internship at ifm ecomatic, building a platform-independent API for embedded mobile controllers (RS232, CAN).

Teaching — Applied Machine Learning

I designed and teach the Master's course Applied Machine Learning at the University of Konstanz — a weekly lecture and tutorial running from data modalities and representation learning through transformers, joint embeddings such as CLIP, diffusion models, and retrieval-augmented generation. Over the term, teams of two build the whole pipeline in Python and PyTorch, ending in a multimodal chatbot that retrieves video clips and generates step-by-step instructions for everyday tasks.

Research — Dimensionality Reduction

I study how classical and neural network-based dimensionality reduction techniques can be evaluated, inverted, and compared, so analysts can reason confidently about high-dimensional structure — work that led to methods and frameworks such as DE-VAE, cPro, and FSDS, and to quality measures for data projections. The same approach applies to quality control, outlier detection, and condition monitoring on industrial process data.

Research — Categorical Data

The focus of my dissertation: I develop measures and visual encodings — ParSetgnostics, the Categorical Data Map, extended Parallel Sets — that reveal patterns in mixed and purely categorical datasets, with applications in software vulnerability analysis and communication studies. In production settings the same questions surface as error classes, machine states, and other ordinal or nominal variables.

Updates

February 2026: Applied Machine Learning Lecture

In the upcoming semester (summer term 2026), I will again teach the "Applied Machine Learning" lecture at the University of Konstanz.

August 2025: Applied Machine Learning Lecture

In the upcoming semester (winter term 2025/2026), I will teach the "Applied Machine Learning" lecture at the University of Konstanz. The course will cover practical applications of machine learning techniques, including generative models for multimodal retrieval.

June 2023: EuroVis

Travelling to EuroVis 2023 (Leipzig, Germany) to attend the 2nd International Conference on Quantification in Visual Computing

Publications

Contact and Social Media

Mail
GitHub
LinkedIn
Scholar
ORCID