A0186
Title: Recent advances in normal approximation for U-statistics
Authors: Dennis Leung - University of Melbourne (Australia) [presenting]
Abstract: Due to applications such as uncertainty inference in ensemble-based predictions, there has been a resurgence of interest in normal approximation theory for U-statistics, both for their complete and incomplete variants and in both low-dimensional and high-dimensional settings. An overview of the state-of-the-art Berry-Esseen-type normal error bounds is provided for U-statistics, with a particular focus on their optimality with respect to the dependence on the kernel's degree and moment properties.
