HANDLING MISSING LABELS WITH MULTIPLE IMPUTATION

Summary from the TLDR AI newsletter

This research introduced a multiple imputation method for evaluating classifiers with missing labels, offering accurate and robust predictive distributions even under MNAR conditions.