“…Continuous active ML approaches have also been proposed that have become particularly prominent in the field, that are similar to dynamic ML for IR approaches widely proposed (Sloan & Wang, 2015). This scheme has been used for a wide variety of purposes including Active Ranking (Wallace et al, 2013), Learning to Rank to identify the initial set of documents, classification of documents once identified (Cormack & Grossman, 2016a), or to identify a stopping strategy for article selection (Cormack & Grossman, 2016b;Cormack & Grossman, 2017;Di Nunzio, 2020;Hollmann and Eickhoff, 2017).…”