Pavel LOSKOT (Mathematical methods)
Research Group
Principle Investigator|Mathematical methods

My research is about developing and applying quantitative mathematical methods to understand dynamic behavior of man-made and natural systems from limited observational data under limited assumptions. The methods include inductive and deductive approaches, causal and statistical inferences, parametric and non-parametric statistics, mathematical logic, and Monte Carlo simulations. The ultimate aim is to maximize the information efficiency and robustness of mathematical models to have good explanatory, descriptive and predictive power. The applications include in silico experiment design, decisions under uncertainty, AI reasoning, spatial time series modeling and processing, explainable simulations, distributed sensing, and events detection and localization.    

Key Words: probabilistic modeling, anomaly detection, statistical signal processing, forward and backward modeling, in silico experiment design, topological signals, graph signals, high-dimensional signals, Monte Carlo simulations