Skip to main content
Fig. 3 | Human-centric Computing and Information Sciences

Fig. 3

From: Correction to: Visual analytics for collaborative human-machine confidence in human-centric active learning tasks

Fig. 3

Dimension reduction techniques for exploring the unlabelled sample pool consisting of all training images. For each row, the unlabelled pool is shown (left), the pool with true labels (middle), and the pool with inferred labels from 100 user-labelled samples (right). Results of labelling accuracy from the inferred scheme, and the accuracy achieved on the standard MNIST test dataset are given using Logistic Regression (LR) and Convolutional Neural Network (CNN)

Back to article page