To discover Exoplanet NASA will use Artificial Intelligence


Last week, Nasa and Google have announced that their advanced computer analysis has discovered two new planets around distant stars, including one that is part of the first-star system with as many worlds as Earth’s solar system.

Our solar system now is tied to the most number of planets around a single star, with the recent finding of an eighth planet circling Kepler-90, a Sun-like star 2,545 light-years from Earth. The planet was found in information from NASA’s Kepler Space Telescope.

The recently discovered Kepler-90i, a sizzling hot, rocky planet that encircles its star once every 14.4 days was determined using machine learning from Google. Machine learning is a way to artificial intelligence in which computers “learn.” In this case, computers studied to recognize planets by finding in Kepler data instances where the telescope recorded signs from planets beyond our solar system, identified as exoplanets.

Just as we assumed, exciting findings are hiding in our archived Kepler data, waiting for the right tool or technology to uncover them,” stated Paul Hertz, director of NASA’s Astrophysics Division in Washington. “This conclusion records that our data will be a treasure trove accessible to innovative scientists for years to come.”

The invention came about after researchers Christopher Shallue, and Andrew Vanderburg instructed a computer to learn how to recognize exoplanets in the light readings recorded by Kepler, the minuscule variation in brightness obtained when a planet passed in front of or transited, a star. Encouraged by the way neurons connect in the human brain, this artificial “neural network” sorted by Kepler data and obtained weak transit signals from a previously-missed eighth planet-encircling Kepler-90, in the constellation Draco.

See Also: Official Post and VideoOfficial Post and Video

While machine learning has earlier been used in searches of the Kepler database, this research shows that neural networks are a promising tool for determining some of the weakest signals of faraway worlds. Other planetary systems apparently hold more promise for life than Kepler-90. About 30 percent larger than Earth, Kepler-90i is so close to its star that its average surface temperature is assumed to exceed 800 degrees Fahrenheit, on par with Mercury. Its outermost planet, Kepler-90h, orbits at a comparable distance to its star as Earth does to the Sun.

“The Kepler-90 star system is similar to a mini version of our solar system. You have small planetoids inside and big planets outside, but everything is scrunched in much closer,” stated Vanderburg, a NASA Sagan Postdoctoral Fellow and cosmologist at the University of Texas at Austin.

Shallue, a senior software engineer in Google’s research team Google AI, came up with the concept to implement a neural network to Kepler data. He became engaged in exoplanet findings after studying that astronomy, like other departments of science, is quickly being inundated with data as the technology for data acquisition from space advances.

“In my spare time, I started googling for ‘discovering exoplanets with large data sets’ and found out about the Kepler mission and the enormous data set available,” said Shallue. “Machine learning shines in conditions where there is so much data that people can’t explore it for themselves.”

Kepler’s four-year dataset consists of 35,000 probable planetary signals. Automated tests, and sometimes human eyes, are used to verify the most promising symbols in the data. However, the weakest signals often are missed practising these methods. Shallue and Vanderburg thought there could be more exciting exoplanet findings faintly hiding in the data.

First, they prepared the neural network to recognize transiting exoplanets using a set of 15,000 previously-vetted signals from the Kepler exoplanet catalogue. In the test set, the neural network accurately recognized true planets and false positives 96 percent of the time. Then, with the neural network learned to identify the pattern of a transiting exoplanet, the experts directed their model to hunt for weaker signals in 670 star systems that already had multiple known planets. They believed that multiple-planet systems would be the best places to look for more exoplanets.

“We got lots of false positives of planets, but also potentially more real planetoids,” said Vanderburg. “It’s like filtering through rocks to find jewels. If you have a finer sieve, then you will catch more rocks, but you might catch more gems, as well.”

Kepler-90i was not the only jewel this neural network filtered out. In the Kepler-80 system, they discovered the sixth planetoid. This one, the Kepler-80g, and four of its nearby planets form what is called a resonant chain where their mutual gravity locks planets in a rhythmic orbital dance. The result is an incredibly stable system, comparable to the seven planets in the TRAPPIST-1 system.

Their analysis paper describing these findings has been accepted for publication in The Astronomical Journal. Shallue and Vanderburg design to apply their neural network to Kepler’s full set of more than 150,000 stars. Kepler has generated a unique data set for exoplanet hunting. After gazing at one patch of space for four years, the spacecraft now is operating on an extended mission and switches its field of view every 80 days.

“These results show the enduring value of Kepler’s mission,” said Jessie Dotson, Kepler’s project scientist at NASA’s Ames Research Center in California’s Silicon Valley. “New ways of looking at the data such as this early-stage examination to apply machine learning algorithms which promise to continue to yield significant advances in our understanding of planetary systems around other stars. I’m sure there are more firsts in the data waiting for people to find them.”

Ames runs the Kepler and K2 missions for NASA’s Science Mission Directorate in Washington. NASA’s Jet Propulsion Laboratory in Pasadena, California, led Kepler mission development. Ball Aerospace & Technologies Corporation operates the flight system with support from the Laboratory for Atmospheric and Space Physics at the University of Colorado in Boulder. This work was performed through the Carl Sagan Postdoctoral Fellowship Program executed by the NASA Exoplanet Science Institute.

About the author

Kanishk Singh

Kanishk Singh, co-founder, and editor-in-chief at The TeCake, has forayed in the Science and Space for over five years, he enjoys his stint as an editor of several local magazines. He has written several editorials and high-level documentations.

Add Comment

Click here to post a comment

You Might Also Like