Detecting COVID-19 Disinformation in News and Social Media

The development work under this 20.R SBIR would involve enhancements to engineer custom COVID-19 related algorithms, add additional data sources and further modifications to the user-interface to surface the additional insights and visualizations.  Primer intends to build upon the commercial solution described above and the body of development work that was provisionally awarded as part of the AFWERX 19.3 D2P2 Phase II SBIR in collaboration with SOCOM, SOCPAC and SOCCENT.  While the 19.3 work will set the foundation for detecting disinformation claims, training COVID-19 specific models for this SBIR will improve the performance (i.e. precision and recall) and increase the speed to detect emerging narratives unique to the global pandemic.  Given the availability of authoritative sources for true claims related to COVID-19, the Primer Platform would be integrated with additional data sources to validate the machine-learning enabled disinformation claims analysis.  In addition, since a significant volume of COVID-19 disinformation claims relate to scientific studies, integrating academic publications may provide a useful supplement to News and Social Media sources.  With these new models and data sources, the user interface will be modified to enable additional visualization and exploration for the COVID-19 use case.