Kepler spacecraft.
Credit: NASA Ames/JPL-Caltech/T Pyle

Researchers using a new technique have added several hundred new exoplanets to create an uptick in the total exoplanet tally.

A team of scientists from Universities Space Research Association (USRA), NASA, and other institutions used a new deep neural network dubbed ExoMiner.

The result: discovery of 301 new exoplanets, joining the 4,569 already validated planets orbiting a multitude of distant stars.

Kepler archive

Leveraging NASA’s Supercomputer, Pleiades, ExoMiner validated the 301 planets using data from the remaining set of possible planets – or candidates – in the Kepler Archive.

Pleiades, one of the world’s most powerful supercomputers, named after the astronomical open star cluster of the same name.
Credit: NASA

NASA’s Kepler spacecraft was lofted in March 2009. Using special detectors similar to those used in digital cameras, Kepler looked for a slight dimming in the stars as planets pass between the stars and Kepler. That mission left a legacy of more than 2,600 planet discoveries from outside our solar system, many of which could be promising places for life.

Highly accurate

“When ExoMiner says something is a planet, you can be sure it’s a planet,” said Hamed Valizadegan, ExoMiner project lead and machine learning manager with the Universities Space Research Association at NASA’s Ames Research Center in Silicon Valley. 

“ExoMiner is highly accurate and in some ways more reliable than both existing machine classifiers and the human experts it’s meant to emulate because of the biases that come with human labeling,” Valizadegan said.

Valizadegan is also the lead author of the paper – “ExoMiner: A Highly Accurate and Explainable Deep Learning Classifier that Validates 301 New Exoplanets” — published in the Astrophysical Journal.

Illustration of NASA’s Transiting Exoplanet Survey Satellite (TESS).
Credit: NASA

Room to grow

“Now that we’ve trained ExoMiner using Kepler data, with a little fine-tuning, we can transfer that learning to other missions,” said Valizadegan. That includes NASA’s Transiting Exoplanet Survey Satellite, or TESS. “There’s room to grow,” he added.

The Kepler and TESS missions have generated over 100,000 potential transit signals that must be processed in order to create a catalog of planet candidates.

USRA’s Miguel Saragoca Martinho, the main engineer behind implementing ExoMiner, pointed out in a USRA statement: “The modular design of ExoMiner allows us to explain why it says something is planet or false positive. That is a peace of mind for domain experts when using a black-box machine classifier such as ExoMiner.”

By utilizing ExoMiner, researchers can distinguish real exoplanets from different types of imposters, or “false positives.” Its design is inspired by various tests and properties human experts use to confirm new exoplanets. And it learns by using past confirmed exoplanets and false positive cases.

For more information on ExoMiner, go to:

Leave a Reply