Scientists are asking space enthusiasts to help train a new artificial intelligence system to unlock the deep secrets of spiral galaxies.
Scientists heading the "Spiral Graph: Cluster Buster" project on the Zooniverse platform have launched a new crowdsourcing campaign funded by a NASA Citizen Science Seed Funding grant.
Led by Dr. Patrick Treuthardt of the North Carolina Museum of Natural Sciences, the project asks everyday volunteers to review automated, color-coded maps of spiral galaxy arms.
When viewed from the side, Treuthardt said a galaxy resembles a flat DVD—exceptionally thin. But looking at it face-on reveals the complex, swirling structures we are familiar with. “Pretty much all the stars you see, all the dust, all the gas, all that stuff is contained within that thin disk,” he explained. “But if you rotate it so that way you look at it face on, then all of a sudden you see all the structure. All that material forms a lot of structure. And so, you see essentially arms that come out from the center of the galaxy that look like, well, the technical term is more like logarithmic spirals. So, they're very much like spirals you kind of see in seashells, things like that.”
Treuthardt said the work builds on a previous project, Spiral Graph, which generated thousands of human tracings that now need to be separated. He said, “We needed to identify individual arms in the galaxy rather than the entire structure as a whole. And there was no good way to disentangle all those tracings into individual groups that signify individual arms.”
By answering a simple yes-or-no question on whether a computer algorithm grouped those arms correctly, human volunteers are providing the vital data needed to teach an A.I. how to recognize patterns. “Essentially, what these people are doing in the Cluster Buster project are helping us create a data set that we can use to train an AI. So, eventually what we want to be able to do is show the AI the set of tracing images and say, ‘Okay, what are the best parameters to group all these traces together into individual arms? Give us those parameters so it works correctly,’” Treuthardt said.
Measuring the exact tightness, or pitch angle, of individual galaxy arms acts as a crucial shortcut. It allows researchers to quickly estimate a galaxy's total mass, its dark matter content, and the size of the black hole at its center. He said, “Just by looking at the spiral structure and kind of measuring that, measuring how tightly wrapped the arms are, that can give you kind of a first estimate as to what the mass of the black hole might be. Would it be very big or would it be a little bit smaller? And so if you're looking, for instance, (for) very massive black holes or very small black holes, you can just do that by looking at the winding of the arms and say, ‘Okay, these likely aren't the ones we're looking for.’”
Related content: Weird Science: Life on a distant planet? Not so fast...
Public interest in the project is already high. Treuthardt said nearly 500 volunteers signed up to participate within the first 24 hours of the project launching.
Best of all, no advanced degree is required. Treuthardt said, “You don't have to be an expert in astronomy at all. You just need an attention to detail and I guess some patience as well. The project itself, I mean, you're just looking at images and you're looking at the detail in these tracing images, right? And you're just looking and verifying whether or not it looks like these tracings are grouped together. So, it takes a little bit of an attention to detail, but you absolutely don't have to be an expert to do this.”
Related content: Countdown to the eclipse; what we'll see in ENC in April and how to see it safely
While the rapid rise of A.I. remains polarizing across the global scientific community due to concerns over data integrity, many astronomers are embracing it. Treuthardt notes that peer feedback shows researchers are eager to let technology handle the heavy lifting. He said, “They feel that it's probably more likely that they will be able to do more astronomy now rather than having to do data science and coding. So, AI could probably do the data science and coding for them, whereas they could just get into the actual astronomy now rather than having to spend a lot of time on that stuff. And I could see that. I could see that potentially being a good benefit for astronomy.”
Though some critics still worry A.I. could narrow human thought or threaten scientific accuracy, the ultimate goal of this project is deeply human. By mapping these celestial coordinates, researchers hope to hunt down rare, intermediate-mass black holes, finally explaining how galaxies like our own Milky Way evolved over billions of years.
