If you had to track down fugitives hidden in five cities around the world, would one day and a $5,000 reward be enough to succeed? And if so, how?
That's what the people behind the TAG Challenge want to know--and what the whole world will soon find out.
On March 31, mug shots of five "suspects" will be published, and it'll be game on in a global hunt for "jewel thieves" in Bratislava, Slovakia; Stockholm; London; Washington, D.C.; and New York City, each of whom will spend 12 hours that day in public areas. The first team to upload photographs of each of the five by noon eastern time on April 1 will win the competition--and with it, a ton of international glory.
Then again, there's a good chance no one will win, given the limited time available to contestants.
The competition is based in part on 2009's DARPA Red Balloon Challenge--in which DARPA hid 10 balloons around the United States and offered a $40,000 prize to the first team that could find them all in a single day.
The TAG Challenge is about learning whether it's possible to implement something similar, said lead organizer Joshua deLara, but in a way that has functional meaning to law enforcement and international security agencies interested in seeing how social media could be used to track fugitives or missing persons.
In 2009, a team from the MIT Media Lab won the DARPA contest with what could be termed a reverse pyramid scheme. Essentially, that team launched a Web site aiming to attract people who themselves knew the balloons' locations or who could help recruit others that might know the proper coordinates, creating a chain effect. When it won, the MIT team spread the reward money down that chain.
With the TAG Challenge, the basic structure is much the same as with DARPA's balloon competition, except that this time, the targets will be on the move, and are located in four different countries. So will the same methods be used to win?
"I don't want to speculate too much about how teams might mobilize or organize or develop their search strategies," said deLara. "But I imagine that some of the same strategies might apply."
For at least one team, that will definitely be the case. Known as CrowdScanner, the team is being led by Manuel Cebrien, a computer science and engineering research scientist at the University of California, San Diego, who was part of the MIT group that won DARPA's balloon challenge.
Cebrien told CNET that CrowdScanner plans on employing a very similar strategy of seeding a network through its Web site and, if victorious, paying out the entire reward down the chain. But since none of the money would end up in team organizers' pockets, why would he bother trying to repeat a three-year-old strategy?
"The problem is that computational social science is a very, very recent field," Cebrien said, and while "we understand very well the structure of social networks...what we don't understand [is the] propagation of information in large networks."
For example, he explained, there is a common perception that social networks are largely responsible for the success of 2011's Arab Spring and Occupy Wall Street uprisings. But in reality, no one knows for sure if that's true. And why? Because there's not enough data to prove the theory, Cebrien argued.
And while a global hunt for five volunteer "suspects" is a very different problem on the surface, he theorizes that the harnessing of social networks in challenges like DARPA's and deLara's could help prove the point. "We need to know if it's true that social networks can allow us to do something we wouldn't be able to do otherwise," Cebrien said, "and the balloon challenge and [the TAG Challenge are] the perfect examples."
Because TAG Challenge organizers aren't requiring contestants to preregister, deLara said he doesn't yet know much about who will be competing, or what strategies will be employed. And he's not at all certain if the participating teams will be able to overcome the time limits and geographic obstacles and identify all five targets.
But he also said he has a side bet with one of his fellow organizers that at least one team will succeed. And if that happens, he believes that there could be a lot to be learned from the winning team's efforts. "If you did ever want to implement this kind of scheme for finding a missing person," de Lara said, "you would want to know something about structuring incentives. So there might be lessons to come out of that."