Propaganda Detection

Propaganda hides in fine-grained language, so detecting it means classifying spans of text rather than whole articles. The page points at one worked example of doing that with BERT.

Medium is Henry Kim's "Fine-Grained Propaganda Detection and Classification with BERT". It starts from the point that propaganda is commonly found in everyday news articles and columns, and that it is dangerous to miss because it shapes information to foster predetermined agendas or ideologies, biasing how readers think about worldwide issues.