Fig. 1.

Top row: stages of the traditional ML framework applied to story recall scoring with minimal human involvement. Bottom row: active learning and human-in-the-loop safeguards which must be incorporated for translational value.