An army of galaxy hunters
Posted: Tuesday, August 07, 2007 6:02 PM by Alan Boyle
More than 85,000 Internet users have signed up to become galaxy inspectors as part of the Sloan Digital Sky Survey's "Galaxy Zoo" project - and you can, too. Inspectors go through a tutorial and click their way through an initial database of 1 million galaxies, classifying them by type. Since the project was launched, less than a month ago, each galaxy in the current database has been checked multiple times (for a total of 12 million checks), but the organizers say there’s much more work that this astronomical army can do.
For nine years, hundreds researchers from around the world have been using the Sloan Digital Sky Survey's telescope in New Mexico to create a three-dimensional map spanning a quarter of the sky. "We're a roadmap or an atlas of the universe," Sloan spokesman Gary Ruderman explained.
GalaxyZoo.org |
Oxford astronomer Caroline Zunckel uses the Galaxy
Zoo Web site to classify a spiral galaxy.
|
The idea behind the Galaxy Zoo is to corral the sky survey's massive files of observations into categories using the wisdom of crowds - or "the wisdom of the public," as Ruderman prefers to put it.
The Galaxy Zoo draws upon the same kind of public interest that has fueled projects such as SETI @ home, which sifts through radio data for patterns suggestive of extraterrestrial intelligence; Einstein @ home, which looks for the signature of cosmic gravitational waves; and Stardust @ home, which has enlisted users to seek out interstellar dust trails.
Astronomers on the Galaxy Zoo team were floored by the response to their plea for help. On the first day the "Zoo" was open, the demand was so great that it overloaded a circuit breaker in the computer room circuit breaker, said Johns Hopkins University's Jan Vandenberg.
"The traffic was 20 times higher than what we hoped for," Johns Hopkins astrophysicist Alex Szalay said in a news release. "This shows the public is really interested in science if they feel they can contribute in a meaningful way."
It's not hard to do: The tutorial guides would-be galaxy hunters through an assortment of objects, and you just have to decide whether each galaxy is elliptical (a fuzzy ball) or spiral (a starry whirl). If it's a spiral, you judge whether it whirls clockwise or counterclockwise. And there's always a "none of the above" category - say, for things that don't look like galaxies, or galaxies that are facing you head-on so you can't tell which way they twirl.
"Computers can do this classification automatically, but humans are far more accurate," said Portsmouth University astronomer Daniel Thomas. "It's like trying to distinguish male and female faces - no computer algorithm will do this as accurately as a person, because we are much better at identifying the most important cues."
Now the Sloan Digital Sky Survey has tens of thousands of eyeballs on the case, instead of mere hundreds. Nearly 7 million images have been checked, and 12.3 galaxy classifications have been registered.
"We now have the world's largest computer working for us, through the combined power of all these human brains," Thomas said.
Ruderman said the goal is to have each galaxy checked at least 20 times, just to make sure there's a consensus for each classification. Particularly interesting anomalies might be flagged for further review, but the main scientific aim is to get at the big picture for galaxy formation: What is the distribution of the different types of galaxies? What determines whether a galaxy becomes a spiral or an elliptical?
"We have theories for how this happens, but to test them we need to know what kinds of galaxies are found in different cosmic environments," Oxford astronomer Anze Slosar said. "The combination of SDSS-II and the Galaxy Zoo will give just the information we need."
But won't the zookeepers eventually run out of things to do? No worries there. Here are the answers to a couple of follow-up questions, e-mailed by Bob Nichol of the Institute of Cosmology and Gravitation at Portsmouth University:
Q: What happens after 20 people classify each of the 1 million galaxies? Is there a "next" project? Where does the Galaxy Zoo go from here?
A: Twenty classifications per galaxy is our first goal, but more classifications per galaxy will only help us more. So there is no formal cut-off, and the more, the merrier. After 20 classifications per galaxy, we really can start using the sample for detailed scientific research.
The project could be expanded in several ways. First, we could simply do the remaining 50 million galaxies in the Sloan Digital Sky Survey! We have only done the tip of the iceberg. These other galaxies are fainter and thus harder to classify. However, we could imagine looking for mergers/interactions in these fainter galaxies, which is easier than looking for spiral arms, for example.
Alternatively, we can use data from the Hubble Space Telescope, which has better resolution and can see distant galaxies better than ground-based telescopes.
Finally, there are many other astronomical data that could be presented to the public in this way - e.g., classification of the spectra of these Sloan galaxies!
Q: When do you think you'll have statistics on galaxy distribution (elliptical vs. spiral)?
A: We can expect the first published results within a year of taking the data. However, I think this database will be important for many years to come - a "famous" database in astronomy.
