Join us on November 6th to continue our Fall 2024 CTML Seminar Series! Wendy Ji's talk "Optimizing Variance Estimation for Causal Inference Through HAL-based Bootstrap" will take place at 11:00AM at Berkeley Way West, 5th Floor, Room 5401.
Variance estimation plays a critical role in accurately inferring causal parameters. Conventional approaches often underestimate variance, potentially leading to elevated type-I errors in limited samples or presence of near-positivity violations. By combining novel variance estimators with a HAL-based bootstrap approach, this project seeks to identify and select optimal variance estimators that balance robustness and computational efficiency. This innovative use of bootstrap methods aims to improve the identifiability of causal estimators and ensure more accurate variance estimation in complex causal inference scenarios.