IARPA selects Raytheon for predictive analytics competition
The Intelligence Advanced Research Projects Activity awarded Raytheon BBN Technologies a $14.5 million contract to develop a system that improves the accuracy of predicting a range of geopolitical events – including elections, conflict, and disease outbreak.
The contract was awarded last year under IARPA’s Hybrid Forecasting Competition, the Waltham, MA-based company announced July 30.
“Event forecasters have struggled with the best way to combine the capabilities of human and machine analysts,” said Lance Ramshaw, principal investigator and lead scientist at Raytheon BBN Technologies. “We’re identifying the optimal combination to develop the best forecasting system possible.”
While human analysts are flexible and can apply context to their event forecasting models, machines can more quickly process massive amounts of data to make unbiased forecasts. Together, Raytheon BBN’s hybrid forecasting system uses the benefits of both by providing:
- Automated aids that increase human forecasters’ efficiency, including intelligent search, automatically computed indicators and alerts about changes or new information.
- Tools for sharing machine reasoning and results with human forecasters.
- Automatic combination of varied human forecasts, with more weight assigned to predictions from historically accurate forecasters.
As prime contractor, Raytheon BBN leads a team that includes Lumenogic and Wright State Research Institute, who contribute to the technology research. Subject matter expertise related to elections and opinion, military events, science, disease outbreak, conflict and leadership and macro-economics are provided by the American Center for Democracy, Ipsos Public Affairs, Systems and Technology Research and Tufts University.