Hi, I’m Chupei, a PhD researcher at the Technical University of Munich (TUM). My work focuses on probabilistic modeling with spatial data. I’m particularly interested in understanding how uncertainty in data affects decisions and how probabilistic methods can make those decisions more transparent and reliable. As part of the FOURIER project, I’m excited to apply probabilistic methods to real-world problems and to collaborate with researchers from various backgrounds for risk management.
This doctoral research focuses on probabilistic modeling for community-level risk management and is carried out at the Technical University of Munich (TUM), Germany. The main objective of the project is to develop probabilistic models capable of analyzing spatial data while explicitly accounting for uncertainties, in order to support informed and reliable decision-making at the community level. The research also aims to identify the key factors that influence model outcomes and to evaluate whether incorporating additional information leads to better and more robust decisions. Special attention is given to efficient methods for capturing spatial and temporal dependencies in data and assessing how these dependencies affect decision-making processes.
The expected outcome of the project is a set of probabilistic models that clearly demonstrate how uncertainties in input data propagate to final results, helping stakeholders better understand and manage risk. For instance, in the context of fire risk assessment, the models can work with approximate inputs—such as estimates of how many people are at home—and still provide practical guidance to firefighters on how to act effectively to minimize potential losses. In addition, the research seeks to support stakeholders in determining whether collecting additional data would meaningfully improve decision quality by quantifying the impact of uncertainty on outcomes. By capturing how information varies across space and time, this work aims to deliver more accurate, actionable insights for community-level risk management.
