Jul 20, 2019 | Updated: 08:54 AM EDT

New Planets Discovered By Artificial Intelligence

Mar 28, 2019 08:11 AM EDT

Close
artificial intelligence helped astronomers discover two new exoplanets
(Photo : NASA | JPL-Caltech | Wendy Stenzel) A new technique in deep learning and artificial intelligence helped astronomers discover two new exoplanets. Traditional planet-hunting methods typically miss potential planets due to data noise, according to a new study.

The researchers from the University of Texas in partnership with Google have utilized the Artificial Intelligence (AI) technology to look into the existence of other possible astronomical objects like a meteor, an asteroid or maybe even a new planet. Through the help of the Kepler space telescope, they were able to discover two new planets and the results are amazing.

The technique of using AI in research proves to be promising as these two new planets would have been difficult to discover if only traditional methods of research were used. The Kepler's extended mission also called K2, have brought into attention possibly two new planets that may exist in the same galaxy as the Earth.

Anne Dattilo, an undergraduate researcher from the UT Austin, has led the research. She created an algorithm that helps sift through data. The data used was gathered through the Kepler. What the algorithm does is that it sifts through all the data to pinpoint possible information that may have been missed out by traditional research methods. The goal is for astronomers to find more useful data from the Kepler. The results of the study have been accepted and will be ready for publication in the next issue of The Astronomical Journal.

NASA Sagan fellow at University of Texas (UT) Austin Andrew Vanderburg and Google Engineer Christopher Shallue are two of the other members of the research team. They both used the Kepler and AI to help uncover the existence of a planet around the Kepler Star. this one was already known to harbor seven other plants in existence. The discovery only helps make that solar system have as many planets like that of our own.

Dattilo explained, "This research project required a different kind of algorithm to bring about new data. If traditional methods were still used, it would have been impossible to uncover a planet or two new ones at that."

The decision to use a different kind of algorithm came about only after the team experienced a mechanical failure. While the planners of the mission were trying to look for something to work around it, the spacecraft had a wobble that needed to be taken into consideration. The K2 mission and the Kepler have identified the existence of thousands of planets around the stars and each one is only waiting for a confirmation.

So why use the AI to help discover more about the space? "AI will help make a uniform search of all the data, something that no traditional research work could do. Because data can sometimes be too noisy, it would be difficult to identify a planet as big as the Earth even if it were just near a star," Vanderburg said.

"There may still be a lot of planets out there and if continue to stick the traditional methods of research, we might end up not discovering anything new anymore," he added further.

©2017 ScienceTimes.com All rights reserved. Do not reproduce without permission. The window to the world of science times.
Real Time Analytics
<