Weapon use in the U.s., new AI system detects the frame of journalistic reports from images and headlines

Weapon use is a big plague in particular in U.s., causing many accidental and intentiously deaths every year. News and information frameworks are crucial in this context. NLP analysis can help to understand how gun violence in the U.s. is conveyd in the news and how this is shaping social cognition about weapon use and possess. The study Detecting Frames in News Headlines and Lead Images in U.S. Gun Violence Coverage by Isidora Chara Tourni, Taufiq Daryanto, Fabian Zhafransyah, Lei Guo, Edward Edberg Halim, Mona Jalal, Boqi Chen, Sha Lai, Hengchang Hu, Prakash Ishwar, Margrit Betke and Derry Tanti Wijaya is a joint work of the Department of Computer Science at Boston University and the Institut Teknologi Bandung in Indonesia. The work has been presented in the context of FEVER – The 4th Workshop on Fact Extraction and Verification in the last 3 days of EMNLP 2021 for the section Finding Papers – NLP Applications.

The study offers the first multimodal news framing dataset related to gun violence in the U.S., curated and annotated by communication researchers. Starting point of the research is the consideration that news media structure their reporting of events or issues using certain perspectives, which in communication research are called “frames”. The research is the first study about the value of combining lead images and their contextual information with text to identify the frame of a given news article. The findings of the article show that using multiple modes of information (article- and image-derived features) improves prediction of news frames over any single mode of information when the images are relevant to the frames of the headlines. “The dataset will allow researchers to further examine the use of multiple information modalities for studying media framing” authors say.