New accepted UKP paper at the Computational Linguistics Journal

2022/01/11

Our paper Annotation Curricula to Implicitly Train Non-Expert Annotators (Ji-Ung Lee, Jan-Christoph Klie, Iryna Gurevych) was accepted for publication at the Computational Linguistics journal.

Our paper Annotation Curricula to Implicitly Train Non-Expert Annotators (Ji-Ung Lee, Jan-Christoph Klie, Iryna Gurevych) was accepted for publication at the Computational Linguistics journal. In this work, we show that ordering instances according to a simple curriculum such as the BERT MLM score can already lead to a significant reduction in overall annotation time.

We sincerely thank our anonymous reviewers and editor for their constructive reviews and patience throughout multiple rounds!

Code and Data are already available on GitHub under an Open Source License. The work itself will be presented at ACL2022.