Pemelajaran semi terawasi

Also known as weak supervision

machine learning approach where noisy, limited, or imprecise sources are used to provide supervision signal for labeling large amounts of training data in a supervised learning setting