We may have just gone one step closer to finding the solution thanks to scientists. The Search for Extraterrestrial Life Using AI Techniques.
By employing a novel algorithm to categorize the data from their telescopes, the team, led by experts from the University of Toronto, has expedited the hunt for extraterrestrial life. This allows them to differentiate between true signals and interference. This has made it possible for them to swiftly analyze the data and identify trends using a technique known as machine learning.
Because it is anticipated that an advanced alien civilization would be advanced enough to send technologically-generated signals, the search for other intelligent life in the cosmos entails finding these signals. Astronomers working on “SETI” (the Search for Extraterrestrial Intelligence) have been searching thousands of stars and hundreds of galaxies for these techno signatures since the 1960s using huge radio telescopes.
Even if the telescopes used for these searches are situated in places where technology, such as mobile phones and TV stations, causes the least amount of interference, human interference still creates significant difficulties. According to Peter Ma, an undergraduate researcher and student at the University of Toronto, “in many of our findings, there is a lot of interference. Additionally, he is the first author of the study that describes this most recent method and was just published in Nature Astronomy. We must separate the fascinating radio waves from Earth from the boring radio signals in space.
The team has taught their machine-learning systems to distinguish between signals that resemble alien life and interference from human activity by replicating signals of both kinds. They evaluated a variety of machine-learning algorithms, compared them, looked at their accuracy and false-positive rates, and then utilized that data to choose a potent algorithm developed by Ma.
Eight additional radio signals that could represent broadcasts from alien intelligence have been found thanks to this new technique. The five separate stars that produced the eight signals were between 30 and 90 light years distant from Earth. A prior examination of the same data that did not employ machine learning missed these indications.
These signals are noteworthy to the SETI team for two reasons. They are different from local interference in that they are there when we gaze at the star and disappear when we turn away, according to Dr. Steve Croft, Project Scientist for Breakthrough Listen on the Green Bank Telescope. Local interference is often constantly present. Second, the signals’ frequency changes over time, making them look far away from the telescope.
The Search for Extraterrestrial Life Using AI Techniques
The two features he describes above can occasionally occur in signals with a dataset of millions of signals, according to Croft, and this can happen purely by coincidence. It’s similar to crossing a gravel path and discovering a stone that seems to fit exactly in the tread of your shoe.
Because of this, despite the fact that the signals resemble what the team anticipates alien signals to look like, the researchers are not yet persuaded that they are from extraterrestrial intelligence — at least not until they encounter the same signal again. The Green Bank Radio Telescope was used to conduct a few quick follow-up observations, but no patterns that may have pointed to alien transmissions were discovered. More research and observations are being conducted.
A describes the method he developed as a synthesis of the machine learning subtypes of supervised learning and unsupervised learning. His method, known as “semi-unsupervised learning,” combines supervised learning approaches to train the algorithm and help it generalize with unsupervised learning techniques to make it easier to find new hidden patterns in the data. In a computer science lesson for grade 12, Ma originally had the notion to use this particular method to look for alien intelligence. Unfortunately, the project baffled his professors, who weren’t sure how to use it.
“I didn’t disclose to my team that this was all an initial high school project that my professors weren’t very fond of until after the article was published.”
New ideas are crucial in a discipline like SETI, according to Dr. Cherry Ng, a research associate at the Dunlap Institute for Astronomy and Astrophysics at the University of Toronto and the paper’s second author. “We might be able to find interesting signals by prodding the data with every approach,” the researcher said.
Machine learning is the way to go in the present era of big data astronomy, according to Ng, who has been working on this project alongside Ma since the summer of 2020. The effectiveness of this strategy in the hunt for alien intelligence has amazed me.
“I’m optimistic that we’ll be able to better assess the probability of alien communications from other civilizations being there with the use of artificial intelligence.”
Ma, Ng, and the rest of the SETI team intend to develop their new method and use it with other datasets and observatories in the future.
Ma claims that the team intends to further scale its machine learning technique using powerful, multi-antenna radio telescopes like MeerKAT, the Square Kilometre Array, and the Next Generation VLA.
“We expect that machine learning can take us from exploring hundreds of stars to searching millions with our new approach and the next generation of telescopes.”