The idea of crowdsourcing scientific research is a powerful one. When a problem is big — really, really big — why not enlist an equally big army of people to solve it? For several years now, the citizen science website zooniverse.org has been hosting projects that let people with no training at all help search for alien radio signals, categorize galaxies, classify bat squeaks and even discover planets orbiting distant stars.
(Photos: Window on Infinity: Pictures from Space)
But for those with a more technical background and a competitive streak, a new science-crowdsourcing initiative — or more accurately, a contest — called “Observing Dark Worlds” is likely to be more appealing. For one thing, it gives participants the chance to be part of truly groundbreaking research rather than simply putting things into categories for the real scientists to study, as some of the projects do. For another, it could earn you $12,000. It could even land you a very lucrative job. And all you have to do is write an algorithm that beats anything highly trained astrophysicists have been able to come up with. Simple, right?
The basic problem that needs to be solved is familiar to anyone who’s even casually interested in the universe. Since the 1970’s, astronomers have known that most galaxies, including the Milky Way, are surrounded by vast clouds, or haloes, of invisible dark matter. We know the role the dark matter plays in the universe: it exerts a gravitational force that helps prevent outwardly expanding galaxies from flying apart. But so far, nobody has figured out what the stuff is made of.
There are ways to go after that question, however. It’s the sheer quantity of dark matter — it outweighs visible stars and gas clouds by five to one or more — that gives it such gravitational muscle. That power is enough not just to help corral galaxies, but to distort light waves as they pass by, which distorts the images of galaxies in the background too. By studying the distortion, astronomers can reconstruct the size and shape of dark-matter haloes in the foreground, the same way you might infer the curvature of a funhouse mirror by noting how it twists the familiar image of your face. And once you do understand the halo, you can use that information to help figure out what the dark matter clouds are made of.
That’s the theory, anyway. In practice, says Thomas Kitching, an astrophysicist at the University of Edinburgh, “The algorithms we have can only pinpoint and characterize clumps of dark matter in a general way. We need orders of magnitude better precision to understand what it’s made of.” So he and a colleague went to a company called Kaggle, which specializes in running competitions to solve thorny problems in computer modeling.
Most of the time, the contests Kaggle conducts are sponsored by companies that have their own data-analysis puzzles that need solving, which can be lucrative for contestants, but not terribly exciting. But now and again the problems are purely scientific. You still need a sponsoring company to bankroll the thing, and in this case, Kitching and his colleagues persuaded a high-tech financial firm called Winton Capital Management to write the check. Winton’s interest in innovative algorithms is academic, yes, but commercial too: anyone smart enough to solve the dark-matter problem is probably also smart enough to write software that will wring a few extra percentage points out of financial markets.
So Winton put up the prize money and programmers around the world have signed up to go after it. “It’s a win-win,” says Kitching. “From the astro side, we were under-resourced. Until now, there were maybe five people in the world who thought about this problem. Now there are 200.”
From Winton’s perspective, laying out a total of $20,000 in prize money (there’s a payout for first, second and third place) is peanuts if they can get the company’s name out to scores of creative programmers from all around the world. And as the broker that sets up and judges contests like this for a fee, Kaggle wins too.
For hotshot programmers who think they can solve the problem, there’s no time to waste. Astronomers have been struggling with the mystery of dark matter for decades. But the Observing Dark Worlds contest will end on December 16th — still time to earn some cash and, oh yes, be etched into science history books forever.
(From the Magazine: The Cathedral of Science)