Scientists identify 50 new planets from old NASA data using artificial intelligence for the first time

Scientists have identified 50 new potential planets with the help of Artificial Intelligence (AI). It is for the first time researchers have used and succeeded in analysing a sample of potential planets and determine which ones are real and which one ‘fakes’.

These planets range from worlds as large as Neptune to those smaller than Earth and with orbits that range from as long as 200 days to as little as a single day.

The astronomers and scientists of the University of Warwick calculated the probability of each candidate to be a true planet.

[caption id=“attachment_8763701” align=“alignnone” width=“1280”] Artist’s rendering of potentially habitable exoplanets, plus Earth and Mars.Image credit: PHL@UPR Arecibo (phl.upr.edu), ESA/Hubble, NASA.[/caption]

A machine learning-based algorithm has been trained to recognise real planets using two large samples of confirmed planets and false positives from now-retired NASA’s Kepler mission.

More from science

Careless Whispers: Scientists discover that plants can 'talk' to each other using smells Life on Mars? European Space Agency’s Orbiter discovers ice water deposits at equator

After this, the researchers used the algorithm on a dataset of the prevailing unconfirmed planetary candidates from Kepler.

Speaking about the finding, Dr David Armstrong from the University of Warwick Department of Physics said in terms of planet validation, no-one has used a machine learning technique before.

He further said that machine learning has been used for ranking planetary candidates, but never in a probabilistic framework, which is what one needs to truly validate a planet.

“Rather than saying which candidates are more likely to be planets, we can now say what the precise statistical likelihood is. Where there is less than a one percent chance of a candidate being a false positive, it is considered a validated planet,” Dr Armstrong said.

He said that almost 30 percent of the known planets to date have been validated using just one method, and that’s not ideal. Dr Armstrong added that while developing new methods for validation is desirable, machine learning lets then do it very quickly and helps prioritise candidates much faster.

The researchers and scientists of the university hope to apply the technique to large samples of candidates from existing as well as upcoming missions including TESS and PLATO, Dr Armstrong said.