IARPA tries to predict the unpredictable
The agency is seeking information on methods for forecasting rare events and alternative ways of assessing such methods.
WHAT: A request for information from the Intelligence Advanced Research Projects Activity on methods for forecasting and modeling rare events
WHY: It is the intelligence community's job to predict the unpredictable and IARPA's job to make the investments necessary for such clairvoyance.
The kinds of events IARPA wants help in predicting include coups, pandemics, nuclear incidents, market crashes, natural disasters and breakthrough discoveries. Officials said they are also looking for alternative ways of assessing forecast methods.
As the RFI states, "The low frequency of rare events makes model and forecast evaluation difficult, as the number of observations may be too small to make conventional statistical inferences about performance." One of IARPA's key forecasting programs -- Aggregative Contingent Estimation -- crowdsources analyses to try to predict world events.
Click here to read the RFI.