In 2011, IARPA announced that it would accept proposals from researchers interested in working on its new Open Source Indicators project, which is designed to trawl through public data, including social media posting like Twitter, to help predict “significant societal events,” such as political protests or riots. “Some of these changes may be indirectly observable from publicly available data, such as web search queries, blogs, micro-blogs, Internet traffic, financial markets, traffic webcams, Wikipedia edits, and many others,” says the agency’s description of the program. The project, which is expected to last up to three years, draws on fields, such as medicine, which have used publicly available data to predict disease outbreaks. For purposes of the project, researchers will look at Latin America because of the availability of public data. It’s expected that the project will draw form a variety of fields, such as the social sciences, mathematics, statistics, computer science and information theory, among others. Researchers selected to for funding will deliver predictions to IARPA and will be graded on how will they are able to “beat the news.”
Source(s): IARPA, Nature, New York Times
Most intelligence estimates are performed by experts, or group of experts. But could the wisdom of crowds prove better at forecasting political events, like how long North Korea’s leader will stay in power, than an elite group of experts? Testing this idea of crowdsourcing intelligence is the underlying principle of the Aggregative Contingent Estimation (ACE) project, sponsored by the Intelligence Advanced Research Projects Activity (IARPA). First announced in 2010, ACE is designed to aggregate individual judgments to help come up with more accurate predictions of future events. “Empirical research outside the intelligence community has shown that the accuracy of judgment-based forecasts is consistently improved by mathematically aggregating many independent judgments,” says IARPA. Crowdsourcing is already used in other areas, particularly with economic forecasting; ACE will test whether such aggregates wisdom could also help improve intelligence. “The goal is to demonstrate the effectiveness of combining the knowledge of many individuals in a unique way that improves accuracy beyond what any one person or expert could provide,” said Charles Twardy, a research assistant professor at George Mason University, which is the heading a research team selected to work on the project. ACE may also eventually use social media to allow participants to discuss their predictions with each other
Source(s): IARPA, George Mason University, IARPA ACE Mason Team