Data-Centric Explanations: Explaining Training Data of Machine Learning Systems to Promote Transparency

Despite the fundamental impact of training datasets on the performance of machine learning systems, information on training data is rarely communicated to the stakeholders. In this project, we explore the concept of data-centric explanations for machine learning systems that describe the training data to end-users. We design explanation prototypes to investigate the potential utility of such an approach and then investigate the reaction of end-users to our data-centric explanations.