A recent analysis of previously recorded satellite data has added 301 new exoplanets to the continuously growing list. This latest addition comes as a result of a new deep neural network tool employed by NASA researchers that can distinguish the signal of an exoplanet better than other methods, possibly even humans.
BACKGROUND: EXOPLANETS LIST KEEPS GROWING
Before 1995, no planets had been spotted outside of our solar system. Since that first discovery, a series of missions by NASA and other research organizations organizations has added to the list, with a recent count marking that total at4,569 “already validated” exoplanets.
Now, thanks to an advanced machine learning tool named ExoMiner, 301 new planets have been added to that list, inching it even closer to 5,000.
ANALYSIS: DEEP NEURAL NETWORK
“ExoMiner is a new deep neural network that leverages NASA’s Pleiades supercomputer, and can distinguish real exoplanets from different types of imposters, or “false positives,” explainsthe press release announcing the newly confirmed planets. “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.”
This tool, the press release also explains, is designed to complement the human researchers manually combing through data recorded during the two Kepler planet hunting missions, and not to replace them. At least, not yet.
“A planet is “confirmed,” when different observation techniques reveal features that can only be explained by a planet,” the press release explains. “A planet is “validated” using statistics – meaning how likely or unlikely it is to be a planet based on the data.”
ExoMiner has significantly more processing power at its fingertips than humans, but the researchers made sure to point out that there is nothing magical going on, just smart science.
“Unlike other exoplanet-detecting machine learning programs, ExoMiner isn’t a black box – there is no mystery as to why it decides something is a planet or not,” said Jon Jenkins, exoplanet scientist atNASA’s Ames Research Center. “We can easily explain which features in the data lead ExoMiner to reject or confirm a planet.”
OUTLOOK: MORE EXOPLANETS TO COME?
The team made sure to note that the planets looked at by ExoMiner had already been marked as viable candidates by Kepler, but were unable to be confirmed until now. They also point out that none of the 301 planets lie in their star’s habitable zones, or are rocky planets like Earth. Still, the team notes, ExoMiner has proven particularly adept at its task, leading researchers to place a high level of confidence in its current, and any future findings.
“When ExoMiner says something is a planet, you can be sure it’s a planet,” sai Hamed Valizadegan, ExoMiner project lead and machine learning manager with the Universities Space Research Association at Ames. “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.”
The team also highlights how this increased ability may be applied to future planet hunting missions, as well as the Transit Exoplanet Survey Satellite (TESS) project currently underway.
“Now that we’ve trained ExoMiner using Kepler data, with a little fine-tuning, we can transfer that learning to other missions, including TESS, which we’re currently working on,” said Valizadegan. “There’s room to grow.”
Follow and connect with author Christopher Plain on Twitter:@plain_fiction