From the hills of West Virginia to the rural plains of Australia, some of the world’s largest telescopes are picking up signals from distant civilizations. The search for extraterrestrial intelligent life, known as SETI, is an effort to find artificial-looking electromagnetic radiation signals that may have come from technologically advanced civilizations in distant solar systems.Research published today1 It describes one of several efforts to use machine learning, a subset of artificial intelligence (AI), to help astronomers quickly sift through the data sets from such surveys. As AI reshapes many fields of science, what can we expect in the search for extraterrestrial life?
Franck Marchis, a planetary astronomer at the SETI Institute in Mountain View, California, said:
The big data problem is relatively new to SETI. For decades, the field has been constrained by a paucity of data. Astronomer Frank Drake pioneered SETI in 1960 when he pointed a telescope at his bank in Greene, Virginia, West, toward two stars and heard a radio transmission. Most of his subsequent SETI searches were also limited to a few stars.
But in 2015, billionaire Yuri Milner funded the largest SETI program ever in Berkeley, California. The Breakthrough Listen project is a project to search his million stars for signs of intelligent life. In this project, West uses telescopes in Virginia, Australia, and South Africa to search for radio emissions that originate from the direction of stars and vary steadily in frequency. This can occur if the alien transmitter is on a planet moving relative to Earth.
blizzard date
The problem is that these searches generate a lot of data. This includes false positives generated by earth interference from cell phones, GPS, and other aspects of modern life.
“The biggest challenge in looking for the SETI signal right now is not getting the data,” says Sophia Shayk, an astronomer at the SETI Institute. “The challenge is to distinguish the kind of signals we’re looking for from human or terrestrial technology and technology elsewhere in the galaxy.”
Making millions of observations manually is not practical. A common alternative approach is to use algorithms that look for signals that match what astronomers think an alien beacon might look like. may miss potentially interesting signals that are slightly different.
Enter machine learning. Machine learning algorithms are very good at removing noise because they can be trained on large amounts of data and learn to recognize features that are characteristic of Earth’s interference.
signal overlooked
Dan Werthimer, a SETI scientist at the University of California, Berkeley, says machine learning is also good at picking up candidates for extraterrestrial signals that don’t fall into traditional categories and may have been missed by previous methods. said.
Peter Ma, a mathematician and physicist at the University of Toronto, Canada, and lead author of today’s paper, agrees. “We can’t always predict what the ETs will send us,” he says.
Ma and his colleagues screened Breakthrough Listen observations of 820 stars made using the 100-meter Robert C. Byrd Green Bank Telescope. They built machine learning software to analyze the data. This yielded nearly 3 million signals of interest, most of which were discarded as earth-based interference. Ma then manually reviewed his over 20,000 signals and narrowed them down to his eight interesting candidates.
The search results ended up being empty. When the team listened again, all eight signals were gone. But the technique can also be used for other data, such as the bulk of observations from his MeerKAT array of 64 radio telescopes in South Africa. This is what Breakthrough Listen started using in his December. Machine-learning algorithms can also be used against his archived SETI data, Ma says, to look for signals that may have gone unnoticed previously.
Citizen SETI
Machine learning is also at the heart of another SETI effort that will launch next month. On February 14th, astronomers at the University of California, Los Angeles (UCLA) will launch a community science project. In this project, civilian volunteers classify images of radio signals, classify them as types of potential interference, and train a machine. – A learning algorithm for retrieving SETI data from Green Bank.
AI can also help other stages of the SETI process. Werthimer and his colleagues use machine learning to rank the stars observed in his ongoing SETI project using his 500-meter FAST radio telescope in China, the world’s largest single-dish telescope. created.
Still, SETI will likely continue to use a mix of traditional and machine learning techniques to classify data, says UCLA astronomer Jean-Luc Margot. Traditional algorithms are still good at detecting candidate signals, and machine learning is “not a panacea,” he says.
This article is reproduced with permission and was first published on January 30, 2023.
