Peter Zaspel

Peter Zaspel has been a W2 Professor of Software for Data-Intensive Applications at the University of Wuppertal since July 2023. Previously, he served as Assistant Professor (March 2022–June 2023) and Acting Professor of Computer Science (Machine Learning) at Jacobs University Bremen gGmbH. His career includes postdoctoral positions at the University of Basel (2017–2019), Heidelberg Institute for Theoretical Studies (HITS) and Interdisciplinary Center for Scientific Computing (IWR) at Heidelberg University (2015–2017), and as a research associate at the Institute for Numerical Simulation, University of Bonn (2009–2015).

Education

• Habilitation in Mathematics, University of Basel, 2019–2021
• PhD in Applied Mathematics, University of Bonn, 2009–2015
• Diploma in Computer Science, University of Bonn, 2004–2009

Research Interests

Peter Zaspel’s research integrates machine learning with high-performance computing and uncertainty quantification. Key themes include multi-fidelity learning, Bayesian inference, kernel-based stochastic collocation, and scalable parallel algorithms for GPUs and distributed-memory systems. His work spans materials science, quantum chemistry, paleoclimate reconstruction, fluid mechanics, and medical imaging.

Projects & Funding

• DFG SPP 2363: “Multi-fidelity, Active Learning Strategies for Exciton Transfer Between Adsorbed Molecules” (2022–2025)
• MarDATA project: “Bayesian chronology modeling for paleoclimate archives” (2022–2025)
• MarDATA project: “Digital ice cores: paleoclimate reconstruction using Bayesian methods” (2022–2025)
• DFG project: “Excitation energy transfer in a photosynthetic system with more than 100 million atoms” (2021–2024)

Invited Presentations

• “Steigerung der Aussagekraft von Vorhersagen durch Unsicherheitsquantifizierung”, WEML2018, Heidelberg, 2018
• “Netzfreie und Multi-Index-Approximationen für parametrische Probleme der realen Welt”, RWTH Aachen, 2018
• “Optimalkomplexitätskernbasierte stochastische Kollokation mit Anwendung in der Strömungsmechanik”, EPFL, 2017
• “Skalierbare Löser für netzlose Methoden auf Many-Core-Clustern”, QUIET 2017, Trieste, 2017
• “H-Matrizen auf Many-Core-Hardware mit Anwendungen in parametrischen PDEs”, University of Kiel, 2016
• “Algorithmische Muster für hierarchische Matrizen auf Vielkernprozessoren”, University of Basel, 2016