Researchers Use AI To Search For Signals From Aliens

JAKARTA - Researchers have successfully used a new algorithm to search for extraterrestrial life based on data from telescopes. In this way, they can distinguish between real signals and interference within a category.

They can also quickly sift through information and find patterns, through an Artificial Intelligence (AI) process known as machine learning.

The study, led by researchers from the University of Toronto aims to find other advanced life in the universe by involving the placement of technologically generated signals or technosignatures.

Because, many assume that advanced extraterrestrial civilizations will be sophisticated enough to transmit that signal. Since the 1960s, astronomers working on the Search for Extraterrestrial Intelligence, or SETI, have used powerful radio telescopes to search thousands of stars and hundreds of galaxies for signs of this technology.

Even though the telescopes used for the search are located in areas where there is minimal interference from technology such as cell phones and TV stations, humans are still a big challenge.

“In our many observations, there are many distractions. We need to distinguish interesting radio signals in outer space from unattractive radio signals from Earth," said the University of Toronto undergraduate student and researcher Peter Ma, who also authored this paper and published it in the journal Nature Astronomy.

The data used in this study came from the Green Bank Telescope in West Virginia, which is one of the main facilities involved in the Breakthrough Listen technosignature search project.

Furthermore, by simulating signals of both types, researchers have trained their machine-learning tools to distinguish between extraterrestrial-like signals and human-generated interference.

They compared a series of different machine learning algorithms, studied their precision and false positive rates, then used the information to solve a robust algorithm, devised by Ma.

This new algorithm has resulted in the discovery of eight new radio signals that have the potential to become transmissions from extraterrestrial intelligence. The eight signals came from five different stars, which are located 30 to 90 light years from Earth.

Those signals were ignored in previous analyzes of the same data, which did not use machine learning. For the SETI team, these signals are important for two reasons.

“First, they are present when we look at the stars and absent when we look away, this is in contrast to local interference, which is generally always present. Second, the frequency of the signal changes over time in such a way that it appears far away from the telescope,” explained Project Scientist for Breakthrough Listening at the Green Bank Telescope, Dr. Steve Croft.

Croft adds, it's important to realize when you have a dataset containing millions of signals, sometimes signals can have both characteristics as described above, simply by sheer coincidence.

"It's like walking across a gravel road and finding a rock stuck in the sole of your shoe that seems to fit perfectly," says Croft.

For this reason, even though the signal appears to be what researchers expect from an extraterrestrial signal, they aren't yet sure it came from extraterrestrial intelligence, at least until they see the same signal again.

When brief observations were made using the Green Bank Radio Telescope, patterns that could indicate extraterrestrial signals were not found. Now, more observations and analysis are being carried out by researchers.

Ma refers to the algorithm he created as a combination of two subtypes of machine learning, namely supervised learning and unsupervised learning.

Called semi-unsupervised learning, his approach involves using supervised techniques to guide and train the algorithm, helping it generalize with unsupervised learning techniques, so that new hidden patterns can be more easily discovered in the data.

Research fellow at the University of Toronto's Dunlap Institute for Astronomy and Astrophysics and co-author on the paper, Cherry Ng says new ideas are especially important in a field like SETI.

“By digging into the data with each technique, we may be able to find interesting signals,” said Ng, who has been working on the project with Ma since summer 2020.

As quoted from the official release of this research, Tuesday, January 31, going forward, Ma, Ng, and the rest of the SETI team hope to extend their new algorithm and apply it to other datasets and observatories.