Goals and Objectives of our research internship?
Aim 1: Interrogate the impact of technical choices during data pre-processing (such as the choice of a missing value imputation method) on stability of the resulting models;
Aim 2: Design pre-processing interventions that improve over-all model stability; and
Aim 3: Quantify disparities in stability between slices that correspond to different demographic groups, and design interventions that lead to stable predictions for all groups.