Research Summaries

Back Discovery of Emergent Groups in Online Media

Fiscal Year 2019
Division Graduate School of Operational & Information Sciences
Department Defense Analysis
Investigator(s) Warren, Timothy C.
Williamson, William
Sponsor Office of Naval Research (Navy)
Summary In rapidly evolving contested information environments, new groups are constantly emerging, forming around images, narratives, and symbols that are fundamental to the competition between groups for resources and influence. Existing ad-hoc approaches have a tendency to miss these dynamics, because they can only search for evidence of groups that are already known to analysts, who are often behind the curve of evolving local patterns of discourse. In contrast, our metrics are derived from unsupervised and partially supervised machine learning algorithms, which allow us to identify patterns associated with emergent radicalized groups not yet known to outside observers.
Hostile information campaigns are frequently aimed at the polarization of local audiences through the promotion of narratives of division along newly politicized dimensions. By developing systematic spatio-temporal maps of the content of human discourse in a region of interest, we can identify patterns associated with polarized audience responses to assess communities at risk of radicalization. Moreover, by combining the measurement of emergent audience radicalization, with the measurement of emergent patterns of information campaigns, we can provide a systematic assessment of the conditions favoring the emergence of symbolic groups in online media, thereby providing key forms of situational awareness for effective maneuver in the information domain.
Our project is organized around a three-year effort. In the first year, we will focus on developing and testing metrics for measuring audience dynamics through publicly available online media content in Singapore and neighboring states. In the second year, we aim to develop underlying software and hardware infrastructure for our next-generation machine learning algorithms, and to present prototype demonstrations of context-aware assessments of emerging groups In the third year, we aim to provide a fully operational system for dynamic situational awareness in the human domain, through the discovery and assessment of emergent groups in online media discourse.
Keywords Machine Learning Information Operations Social Media
Publications Publications, theses (not shown) and data repositories will be added to the portal record when information is available in FAIRS and brought back to the portal
Data Publications, theses (not shown) and data repositories will be added to the portal record when information is available in FAIRS and brought back to the portal