
Johannes Gutenberg University Mainz
Institut für Informatik
Staudingerweg 9
55128 Mainz, Germany
E-Mail: koebschall@uni-mainz.de
Room: 03-623
Research Interests
- Online learning, data streams
- Resource-awareness
- Transparency of AI models
- Stochastic methods
- Deep neural networks
Short Scientific CV
since 07/2022 Research associate at Johannes Gutenberg University, Mainz.
09/2024-12/2024 Guest research associate at Lamarr Institute, TU Dortmund, Germany.
06/2019 - 11/2021 Computer Science (M. Sc.) at Johannes Gutenberg University Mainz, Germany.
01/2019 - 06/2019 Semester abroad at Universitat de València, Spain.
10/2015 - 03/2019 Computer Science (B. Sc.) at Johannes Gutenberg University Mainz, Germany.
Teaching
- Supervision and organization in "Machine Learning Seminar: Real-Time Insights from Streaming Data, Time Series, and Forecasting Models" (WS 25/26)
- Co-Supervision in Data Mining and Machine Learning seminar (WS 22/23, SS 23, WS 23/24, SS 24, WS24/25, SS 25)
- Workshop "KI als Chance oder Risiko?" (SS 24)
- Tutor in Computability and Formal Languages (SS 21)
- Tutor in Complexity Theory (WS 19/20, WS 20/21)
Presentation and Poster
- Paper "Transparent and Adaptive Pruning of Hoeffding Trees" presented (Talk) at 2nd TempXAI Workshop for Explainable AI in Time Series and Data Streams, ECML-PKDD 2025, Porto, Portugal, 15-19 September 2025.
- Talk "Balancing Prediction Performance, Transparency and Energy Consumption in Machine Learning Models for Data Streams" at Fast Machine Learning for Science Conference 2025, ETH Zurich, Switzerland, 1-5 September 2025.
- Poster "Soft Hoeffding Tree: A Transparent and Differentiable Model on Data Streams" at On Boarding-Meeting Trading Off Non-Functional Properties of Machine Learning (TOPML), 29 October 2024.
- Paper "Soft Hoeffding Tree: A Transparent and Differentiable Model on Data Streams" presented (Talk) at Discovery Science 2024, Pisa, Italy, 14-16 October 2024
- "Soft Hoeffding Tree: A Transparent and Differentiable Model on Data Streams" presented (Talk and Poster) at 1st Mainz and Friends Artificial Intelligence Conference (MAInC), Mainz, Germany, 13-14 June 2023.
- "Soft Hoeffding Tree: A Transparent and Differentiable Model for Data Streams" presented (Talk and Poster) at Trading Off Non-Functional Properties of Machine Learning (TOPML) Workshop, 3 February 2023.
Publications
2025
Kobschall, K., Hartung, L., and Kramer, S. (2025). Adaptive differentiable trees for transparent learning on data streams. MACHINE LEARNING, 114(11). DOI Author/Publisher URL
Koebschall, K., Hartung, L., and Kramer, S. (2025). Soft Hoeffding Tree: A Transparent and Differentiable Model on Data Streams (Vols 15243, pp. 167-182). DOI Author/Publisher URL
Further Links
Link to my Github profile.
Link to my LinkedIn profile.
Link to my Google Scholar profile.
