Training data inference quantifies the impact of training samples on a model's predictions. The community focuses on efficient means to estimate the influence of individual or groups of training samples.