Inferring Social and Psychological Meaning in Social Media

While recent events have highlighted the clandestine methods hardened jihadist resort to for their tactical planning for execution (game consoles, encrypted channels, etc.), the initial recruitment and radicalization still occurs via popular online social media outlets, and in particular Twitter. These have still proven to be their most effective tool to reach the broadest susceptible audience responsive to their message. But despite recognizing Twitter as a primary enabler for ISIS marketing and recruitment, finding ISIS supporters and recruiters and disabling them has been ineffective. Current techniques to detect such users generally rely on crowd sourcing or applying natural language processing techniques to parse content. Neither of which is practical on a large scale, especially given variations in language, and content that includes links and videos. The Social Understanding and Reasoning Framework (SURF) finds and classifies social media users using a combination of machine learning and advanced classification algorithms extended from the bioinformatics domain. This Phase II will extend the algorithms and broaden the classification groups to meet additional intelligence requirements for Army.