In 2015, on the same year, a magnificent observatory Soil After securing evidence of the 4D fabric of space-time, scientists began toying with a rather far-fetched idea: If intelligent aliens were out there, could they try to build a scientific megastructure of their own? If so, can we find it? In fact, were we already like that?
Yes, I’m completely serious. It all starts with a fascinating event paper “Planet Hunters IX. KIC 8462852 – where is the flow?” titled
In this article, a team of researchers presented their analysis of data compiled from NASA’s studies. Kepler telescope. It was related to a star that resided for approximately 1,470 years light years KIC 8462852 from your residence or Boyajian’s Star As a nod to the lead author of the study. According to the team’s results, Boyajian’s Star appeared to exhibit some very strange dips in the light.
Normally, when examining a star from our vantage point in the universe, telescopes can naturally see dips in starlight when something passes between it and the star. Imagine looking at a bright light bulb, then someone walks past the light bulb. Their emissions will appear to be interrupted. As you can usually guess outer planet The star causes such dimming as it orbits its host – but… not for Boyajian’s Star.
“This is not a sphere,” said postdoctoral researcher Daniel Giles. SET The institute said this at the 243rd meeting of the American Astronomical Society in January. “It consists of what looks like a bunch of panels… it looks like what a megastructure would look like.”
That’s why the crowd went wild after the 2015 result. News articles, follow-up observations, opinion pieces, and even general conversations began to flood the Internet. astronomy niche OK, pause. I’ll save you the trouble and let you know that the final consensus is: No, these weird drops not welded a giant piece of futuristic alien technology. “It’s probably dust,” Giles said. But here’s the thing.
Relating to: ‘Dyson sphere’ legacy: Freeman Dyson’s wild alien megastructure idea will live on forever
“Signals like this were actually missed in the Kepler data,” Giles explained. In fact, a big reason why the researchers behind the paper found the lightfall anomaly was because citizen scientists noticed it by accident while looking for something else.
Or, as Giles puts it: “People weren’t looking.”
That’s exactly what he and other researchers aim to do. Perhaps they believe the truth about aliens lies directly in the data; We just need to look for it. But like, Really To look.
Registering machines
In short, Giles and his team aim to investigate puzzling, mysterious, intriguing and highly unusual signals in the data collected by NASA’s Transiting Exoplanet Survey Satellite. TESS. They want to hunt for starlight falls that have no defined shape, no defined depth, or even a defined time period. Cosmic outliers.
Strange dips like this can be detected through photometric curves representing brightness. time. “We’re counting photons,” Giles explained simply. But the real kicker is exactly how the team wants to embark on this anomaly-hunting quest: machine learning.
The process is pretty much as follows.
The TESS data used in the study is based on the satellite’s view of different sky sectors. These sectors were imaged for approximately 30 days at a time; During this scan, TESS took a snapshot of the observed area every 30 minutes. This ultimately allowed the team to have nearly 60 million light curves ready for analysis. stars brighter than 14 size. In the magnitude system, smaller numbers are brighter than larger numbers; For example, an object of magnitude 0 is 100 times brighter than an object of magnitude 5. A. full moon It enters the negatives with a magnitude around -12.6; Sun It shines at magnitude -27. And such.
The next step is to start organizing the light curves collectively based on things like their shape and periodicity. “We process 60 million different light curves, so we need them to be cheap and easy to calculate,” Giles said. “We compute these inexpensive metrics and then run anomaly detection on top of that, and it’s an intensity-based anomaly detection; we find what has features that stand out.”
Then, after reducing the data to a manageable size, the team prepares to apply more detailed techniques, which often require more computing power. Subtle, difficult analyzes to make. “We make sure the behavior actually exists, is astrophysical, and is not caused by an instrumentation issue,” Giles said.
If something exhibits a recognizable pattern, it’s time to go back to the sorting phase.
“We’re finally going manual,” Giles said, “because there’s nothing better than the human eye at finding strange things.”
You might need a human to find an alien
Honestly, I was so excited to hear about something that essentially humans could find weird things that no machine could ever find. I think this underpins our frantic efforts to locate intelligent aliens. I think we are curious by nature and somehow drawn to deviations from patterns.
“There is a certain level at which we can use ML methods, but ultimately we need to be able to understand why things are happening,” Giles told Space.com.
Perhaps a pool full of highly accurate data sets is a pool full of highly accurate data sets until a human starts parsing them to make connections that a machine is not yet programmed to recognize.
“There’s an additional trick for things like anomaly detection,” Giles said. “There’s no ground truth, so we can’t necessarily train something to find the weirdest things or the most interesting things because we don’t know exactly what those things will be.”
RELATED STORIES:
— No Alien Massive Structure: Star’s Strange Darkening Likely Caused by Dust
— ‘Alien Megastructure’ Star Strikes Again with Strange Blackout
— Will we be able to communicate with aliens?
Even when it comes to standard robotics, which aims to mimic human structure, a limiting step for scientists is decoding the physical laws that determine the way we move. Because as humans we don’t really need to know how some aspects of humanity work. They just work. A few years ago, for example, a team made major progress in figuring out how our fingerprints affect our ability to grasp. When you wash slippery dishes, you instinctively know how hard it is to hold them so they don’t fall out of your hands. You are unconsciously evaluating your fingerprints throughout this entire process. But scientists literally had to do it. a new law of physics translating that instinct into written reality.
There seems to be a similar concern for machine learning and artificial intelligence, although both can technically be trained to come up with their own solutions. It’s hard to program a machine to find something we couldn’t find before, because what would we tell it to look for? This is like scientists defending the James Webb Space Telescope as an invention that could answer some cosmic questions we never thought to ask.
“There are limits to what AI and machine learning can do for us, but there are also many opportunities as long as we understand what machine learning specifically does,” Giles said.
Food for thought.
But Giles says the team is also trying to look for specific anomalies that can actually be coded. “We injected about 2 million different artificial signals into the light curves, which didn’t have any decay signature that we knew of, but still had artifacts in them, so their behavior continued,” he said.
As for the abnormal results so far? “So far none of these speak to us as if they were megastructures,
“But they’re certainly interesting.”