Fig. 2From: Visual analytics for collaborative human-machine confidence in human-centric active learning tasksActiVAte user interface. Key components include the classifier view for dragging new instances to corresponding classes, sample pool view, which offers a dimensionality reduction view of the unlabelled pool of samples, and the accuracy view that shows the accuracy against the test set at each iteration of training. Configuration options are also available to the user in the left columnBack to article page