A Divide-and-Conquer Approach to Geometric Sampling for Active Learning

Xiaofeng Cao

Active learning (AL) repeatedly trains the classifier with the minimum labeling budget to improve the current classification model. The training process is usually supervised by an…