DH2026 Conference in Daejeon

Research work is not always only about writing code, training models and interpreting results. Behind all these activities, a huge amount of manual work is hidden. That's exactly what Raven and Klara presented at the DH2026 conference in Daejeon, South Korea: Manual annotation efforts that serve as a basis for training automatic annotation models. Thanks to manual annotations and subsequent models training, we can recognize where on the page job ads are, what job titles are mentioned within the ads, which salary is being offered, and far more.