no code implementations • 7 Mar 2019 • Thomas Drugman, Janne Pylkkonen, Reinhard Kneser
The goal of this paper is to simulate the benefits of jointly applying active learning (AL) and semi-supervised training (SST) in a new speech recognition application.