NASA Just Found Our Solar System's Twin by Using AI

Thursday, 14 December 2017 - 5:43PM
Space
Technology
Artificial Intelligence
No
Thursday, 14 December 2017 - 5:43PM
NASA Just Found Our Solar System's Twin by Using AI
Image credit: NASA
Searching the stars for unique phenomena is not an easy process.

The problem is that space is simply too big, too diverse, and too wonderful—locating a specific kind of anomaly among the many wondrous sights scattered throughout the cosmos is near impossible for humans, without easily-distracted brains. With so many stars to check, the process of scanning the galaxy to find planets like our own can take a lot of time and effort.

Thankfully, artificial intelligence can help us in the process of spotting distant stars and their neighboring planets. NASA has announced that, thanks to an AI program that was given the task of spotting cool stuff in space, the agency has been able to find a solar system that looks uncannily like our own; albeit in miniature form.

The Kepler-90 system exists a distant 2,545 light years from Earth, but has drawn attention from the astrological society after an AI noted that its series of eight planets match up well with our own. The primary difference is that its planets orbit a lot closer to the sun than those in our solar system, with the newly discovered Kepler-90i making a full rotation around the star in a matter of just fourteen Earth days.

In order to locate Kepler-90's planets NASA's AI had to scan through a daunting thirty five thousand potential signals from distant stars, over a period of four years.

This is where machine learning was able to come into play to help make the process easier—the AI was fed data from around fifteen thousand signals that NASA had previously investigated, so the AI had a pretty good idea of what it was looking for based on the kinds of readings that NASA had flagged as noteworthy among the program's database of reference materials.



From there, it was a simple matter of letting the AI run checks for all potential star systems against its database until the program found something that matched what it was looking for, which happened to be a bunch of newly discovered planets orbiting Kepler-90.

According to Paul Hertz, director of NASA's Astrophysics Division in Washington:

Opening quote
"Just as we expected, there are exciting discoveries lurking in our archived Kepler data, waiting for the right tool or technology to unearth them. This finding shows that our data will be a treasure trove available to innovative researchers for years to come."
Closing quote


Kepler-90 isn't actually the most exciting solar system in the galaxy—it's unlikely that its superhot worlds will bear life, or even any noteworthy new discoveries.

What is special, is the fact that an AI managed to identify Kepler-90 as fitting the right parameters for investigation. This shows that there really are benefits to employing machine learning as a technique for searching the cosmos for interesting research subjects without the need for a human to slog through thousands of signals in order to find a few interesting stars that warrant a closer look.

This kind of AI can help us to narrow down our search for Earth-like planets, and can make it easier to spot alien life, should any exist anywhere in the universe.

Essentially, NASA is building a self-teaching search engine that can trawl through all of our records of the stars to find things that look interesting, based only on a vague description of what scientists are looking for.

The future of space exploration is going to be a whole lot easier if we can trust an artificial intelligence to do all the boring stuff for us.
Science
NASA
Space
Technology
Artificial Intelligence
NASA Just Found Our Solar System's Twin by Using AI
No
No