Active Learning: Problem Settings and Recent Developments
This paper provides a review of active learning, which is a method to reduce data labeling costs for researchers and practitioners in supervised learning.
This paper reviews active learning, a method to achieve high-precision predictive models with limited labeling costs by adaptively selecting samples for labeling. It covers basic problem settings, recent research trends in acquisition functions, theoretical advancements, and stopping criteria for sequential data acquisition.
In supervised learning, acquiring labeled training data for a predictive model can be very costly, but acquiring a large amount of unlabeled data is often quite easy. Active learning is a method of obtaining predictive models with high precision at a limited cost through the adaptive selection of samples for labeling. This paper explains the basic problem settings of active learning and recent research trends. In particular, research on learning acquisition functions to select samples from the data for labeling, theoretical work on active learning algorithms, and stopping criteria for sequential data acquisition are highlighted. Application examples for material development and measurement are introduced.