Uncertainty Propagation

Uncertainty propagation focuses on accurately estimating the impact of input uncertainties on the outputs of complex systems, models, and algorithms. Current research emphasizes developing efficient and reliable methods for propagating uncertainty through various model architectures, including neural networks (e.g., Bayesian neural networks, Gaussian processes), and employing techniques like moment matching, polynomial chaos expansion, and conformal prediction. This field is crucial for enhancing the trustworthiness and reliability of AI-driven systems in safety-critical applications like aviation and autonomous vehicles, as well as improving the accuracy and robustness of scientific simulations and data analysis.

Papers