IARPA launches Hybrid Forecasting Competition
The Intelligence Advanced Research Projects Activity, in the Office of the Director of National Intelligence, announced on August 18 a multi-year research effort to improve geopolitical forecasting. The Hybrid Forecasting Competition—“HFC”—program aims to improve accuracy in predicting worldwide geopolitical issues, including foreign political elections, interstate conflict, disease outbreaks, and economic indicators by leveraging the relative strengths of humans and machines.
“Policymakers rely on the intelligence community to help them anticipate geopolitical events, and over the years, IARPA has led efforts to improve forecasting accuracy,” said Dr. Seth Goldstein, IARPA HFC program manager. Previous efforts, such as IARPA’s Open Source Indicators—“OSI”—program, have explored the power of automated prediction models. Prediction markets and IARPA’s Aggregative Contingent Estimation—“ACE”—program have harnessed crowd wisdom to make accurate forecasts. “By combining both approaches, we have a strong potential to be more accurate than either method alone while increasing the number of forecast problems that can be addressed,” Dr. Goldstein said. The methods HFC aims to develop should help efficiently and effectively navigate the challenge of combining human judgments with machine model-based forecasts to maximize the accuracy of the resulting forecast.
The HFC program will pit a variety of human-machine hybrids against each other in a series of competitions involving the predictions of real world events by attempting creative fusion of human judgments with information from sources such as traditional news reporting, social media, and financial indicators. Over the course of the four-year program, forecasting questions will become more challenging—increasing in geographic and temporal specificity, and will come at a faster pace—doubling over the course of the program.
Through a competitive Broad Agency Announcement, IARPA has awarded HFC research contracts to teams led by HRL Laboratories, LLC (with subcontracts to Harvard University; University of Houston; Northeastern University; and Massachusetts Institute of Technology), Raytheon BBN Technologies Corporation (with subcontracts to American Center for Democracy; Wright State Applied Research Corporation; Tufts University; System of Systems Analytics, Inc.; Lumenogic, Inc.; and Ipsos Public Affairs, LLC), and the University of Southern California (with subcontracts to Barnard College; Ward Associates; University of California, Irvine; Stanford University; Regina Joseph, Independent Consultant; Pavel Atanasov, Independent Consultant; Manuel Cebrian, Independent Consultant; Kun Zhang, Independent Consultant; and Fordham University). MITRE Corporation, Good Judgment, Inc, and Cultivate Labs will work together to independently test the new systems—drawing on thousands of volunteers to participate as forecasters.