Scholar
Dmitry Kobak
Google Scholar ID: BUQbD5kAAAAJ
University of Tübingen
Citations & Impact
Citations
5,647
H-index
26
i10-index
36
Publications
20
Co-authors
111
list available
Publications
Resume (English only)
Research Experience
- Has taught an introductory machine learning course for MSc students in neuroscience and data science at Tübingen for several years
- Taught a BSc course 'Einführung ins Machinelle Lernen' (in German) and an MSc seminar on 'Transformers, large language models, and their use in physics' at Heidelberg University in winter 2023/24
- Supervises postdoc Sebastian Damrich (since 2023)
- Advises PhD students Rita González Márquez (since 2022) and Niklas Böhm (since 2021)
- Mentored multiple MSc students, several of whom continued as PhD students or moved to institutions like the University of Zürich
Background
- Group leader in the Department of Data Science at the Hertie AI Institute, University of Tübingen, Germany
- Research interests include self-supervised and unsupervised learning, particularly contrastive learning, manifold learning, and dimensionality reduction for 2D visualization of scientific datasets
- Works with image, text, graph, and single-cell RNA-seq data in neuroscience contexts
- Interested in statistical forensics; involved in analyses of Russian electoral falsifications, war fatalities, and Covid-19 excess mortality
- Privatdozent at the Faculty of Computer Science
- Served as Vertretungsprofessor (visiting professor) at Heidelberg University during the 2023/24 winter semester
- Member of the ELLIS Society, the Cluster of Excellence «Machine Learning for Science», and an IMPRS-IS associated scientist
