A lot of controversial information goes viral on social media
This growing phenomenon calls for publicly accessible tools for limiting the spread of misinformation. Moreover, journalists have to rely more and more on user-generated content, many coming from social media, especially for example if they do not have direct access to a region where significant events occur. An investigation needs to be carried out to decide whether a shared content represents real information. In this context, there is a growing interest on developing verification tools to help news professionals, and more generally any user, to assess the veracity of an information posted on social media. This technology is still in its early stage, and usually this task requires to examine different modalities, and provide a variety of different tools.
This special session is dedicated to technologies that could help journalists or social network users
The task is made even more difficult by the diversity of misleading use: digitally manipulated content, reposting of real content in a wrong context, content created from scratch (like synthetic multimedia presenting as real).
The image forensics tools help detecting manipulated content, but the text analysis could also be used to measure the truthfulness of a document, as well as the reputation of the author or relationships between diverse individuals and organization could ease the discovery of fake content. Furthermore, collecting related information on the web is an important source of information.
This special session is dedicated to works addressing the problem of social media verification, assessing the veracity of user-generated content, or providing technologies that could help journalists or users, from any angle: link analysis, image and video forensics, natural Language processing, information propagation analysis, rumor analysis, spatio-temporal entities extraction…