Over the years, Scientists have identified thousands of possible Planets outside our solar system using NASA’s now retired Kepler Telescope. The telescope, over its decade-long observation, gathered information on thousands of potential exoplanets by capturing a distinctive dip in the light from distant stars.
The telescope worked on the assumption that a planet’s passage might be causing the light to decline. To confirm the possibility of a Planet, Scientists have to sift through a lot of data and that can be quite tedious and time consuming.
A team of British researchers led by David Armstrong at the University of Warwick in the UK have identified 50 new planets using artificial intelligence, marking a technological breakthrough in astronomy. The researchers have worked out an efficient way to harness artificial intelligence and machine learning to handle planet confirmation.
“The algorithm we have developed lets us take 50 candidates across the threshold for planet validation, upgrading them to real planets,” Armstrong said in a Warwick release Tuesday. And now that astronomers know the planets are real, they can prioritize them for further observation without wasting their time on “fake” planets.
This is the first time Scientists have successfully leveraged AI and Machine learning tools to validate a planet without any human intervention or validation.
“In terms of planet validation, no-one has used a machine learning technique before,” stated David Armstrong of the University of Warwick, one of the study’s lead authors, in the news release. “Machine learning has been used for ranking planetary candidates but never in a probabilistic framework, which is what you need to truly validate a planet.”
How the AI algorithm found the planets?
The researchers trained the algorithm on two large samples of confirmed planets and false positives from Kepler to get the planet-detection system ready. Using that information, it learned patterns to distinguish between real and fake planets and managed to find the actual batch of 50 – when put to test on a dataset of still unconfirmed planetary candidates from Kepler.
The new planets confirmed by the AI range between Earth to as big as Neptune and had orbits ranging from 200 days to as little as one single day.
“Rather than saying which candidates are more likely to be planets, we can say what the precise statistical likelihood is,” Dr. David Armstrong said. “Where there’s less than a 1% chance of a candidate being a false-positive, it is considered validated planet.”
The algorithm could “validate thousands of unseen candidates in seconds,” the study revealed. And since it’s based on machine learning, it can be improved upon, and can become more effective with each new discovery.
In their study, the research team suggests that astronomers should use multiple validation techniques — including this new algorithm — to confirm future exoplanet discoveries.
Dr. Armstrong adds: “Almost 30% of the known planets to date have been validated using just one method, and that’s not ideal. Developing new methods for validation is desirable for that reason alone. But machine learning also lets us do it very quickly and prioritize candidates much faster. We still have to spend time training the algorithm, but once that is done it becomes much easier to apply it to future candidates.
The technique holds promise for sifting through the large amounts of data collected by projects such as TESS and ESA’s planned PLATO mission. TESS’ primary mission alone spotted 66 new exoplanets and 2,100 candidates by mapping out the sky.