Dr. Paul Tremper

My research focuses on machine learning and computational methods for modelling and understanding complex systems. With a background in theoretical physics (Dr. rer. nat.), I am particularly interested in researching how mathematical structure and domain-specific knowledge can be incorporated into data-driven models of complex systems.

My work spans probabilistic machine learning, Bayesian inference, neural networks and neural operators, with applications in environmental sensing, spatio-temporal modelling, inverse modelling and computational fluid dynamics. I am especially interested in approaches that combine the flexibility of machine learning with the interpretability and structure of mathematical and physical models.

• since 2019 Scientific Researcher at TECO
• 2018 Dr. rer. nat. Dissertation in Theoretical Particle Physics
• 2014 - 2018 PhD student at TTP (Institute for Theoretical Particle Physics) at KIT
• 2014 Graduation with Diplom (Master equivalent) from Karlsruhe Institute of Technology (KIT)
• until 2014 Student of Physics at Karlsruhe Institute of Technology (KIT)

• Scientific Machine Learning: Machine learning methods for modelling, simulating, and understanding complex scientific and physical systems.
• Physics-Informed and Structure-Aware Machine Learning: Incorporating mathematical structure, physical principles, and domain knowledge into machine learning models.
• Probabilistic Modelling and Bayesian Inference: Gaussian processes, uncertainty quantification, statistical inference, and inverse problems.
• Neural Networks and Neural Operators: Deep learning architectures, operator learning, and surrogate models for scientific computing and simulation.
• Environmental and Spatio-Temporal Modelling: Data-driven modelling of environmental processes, sensor networks, atmospheric dispersion, and computational fluid dynamics.