The products and services in this category are used to apply labels to training data for the purpose of training models with what is assumed to be ground-truth. Labeling can be done by humans, machines, or a hybrid of the two. There are also new and emerging techniques, processes, vendors, and products emerging that leverage active learning (where the system or model can specify which training data it needs next in order to improve its accuracy) and techniques such as programmatic labeling.
