Single Shot Active Learning using Pseudo Annotators

Yazhou Yang, Marco Loog

Standard myopic active learning assumes that human annotations are always obtainable whenever new samples are selected. This, however, is unrealistic in many real-world application…