A0867
Title: New root-n consistent, numerically stable higher-order influence function estimators
Authors: Lin Liu - Shanghai Jiao Tong University (China) [presenting]
Abstract: Higher-Order Influence Functions (HOIFs) provide a unified theory for constructing rate-optimal estimators for a large class of low-dimensional (smooth) statistical functionals/parameters (and sometimes even infinite-dimensional functions) that arise in substantive fields, including epidemiology, economics, and the social sciences. Since introducing HOIFs, they have been viewed mainly as a theoretical benchmark rather than a valuable tool for statistical practice. Works aimed to flip the script are scant, but a few recent papers make some partial progress. A fresh attempt at achieving this goal by constructing new, numerically stable HOIF estimators (or sHOIF estimators for short, with s standing for stable) is taken with provable statistical, numerical, and computational guarantees. This new class of sHOIF estimators (up to the 2nd order) was foreshadowed in synthetic experiments conducted by other researchers.
