Our previous work on propaganda has addressed propaganda detection at the document level.
We propose a novel task: performing fine-grained analysis of texts by detecting all fragments that contain propaganda techniques as well as their type. In particular, we create a corpus of news articles manually annotated at the fragment level with eighteen propaganda techniques and we propose a suitable evaluation measure.
We further design a novel multi-granularity neural network, and we show that it outperforms several strong BERT-based baselines.
The code used in the experiments is available here.
Wanna see our system in action? Check our demo, which collects articles on a topic and shows statistics on the propaganda techniques as well as all techniques for every article.