Tianbai Xiao

Tianbai Xiao is a Researcher at the Karlsruhe Institute of Technology (KIT) within the Department of Mathematics and Steinbuch Centre for Computing. His work spans mesoscopic science, uncertainty quantification, and scientific machine learning, focusing on multi-scale, multi-physics problems in flow transport.

His research in kinetic theory addresses nonlinear partial differential equations, hyperbolic conservation laws, and the unified modeling of continuum/rarefied flows. He develops high-performance numerical algorithms like the Unified Gas-Kinetic Scheme (UGKS) and Kinetic.jl (a finite volume toolbox for scientific computing). Current projects include mesoscopic science, stochastic data science, and physics-informed neural networks.

He contributes to open-source tools including FluxReconstruction.jl for advection-diffusion methods and Langevin.jl for stochastic kinetic modeling. Publications cover Journal of Computational Physics, Engineering Fracture Mechanics, and Entropy, with preprints on arXiv in 2025 addressing force-driven flows and hybrid peridynamics.

Teaching activities include the Introduction to Kinetic Theory lecture at KIT, and mentoring in the CAMMP (Computational and Mathematical Modeling Program) to develop problem-solving skills through real-world modeling tasks. He advocates for problem-based learning where students translate non-mathematical problems into mathematical language.