Extra knowledge and machine studying have pushed SETI into overdrive

Astronomers and astrophysicists have been searching for extraterrestrial intelligence (SETI) for over sixty years. This consists of listening to other star systems for signs of technological activity (or “technosignatures”) such as radio transmissions. This first attempt was known as Project Ozma in 1960, in which the famous SETI researcher Dr. Frank Drake (father of the Drake equation) and his colleagues used the Robert C. Byrd Green Bank Telescope in West Virginia to conduct a radio survey of Tau Ceti and Epsilon Eridani.

Since then, the vast majority of SETI surveys have similarly looked for narrow-band radio signals, since they are very good at propagating in interstellar space. However, the biggest challenge has always been filtering out radio transmissions on Earth – aka. Radio Frequency Interference (RFI). In a recent study, an international team led by the Dunlap Institute for Astronomy and Astrophysics (DIAA) applied a new deep learning algorithm to data collected by the Green Bank Telescope (GBT), revealing eight promising signals, that will be of interest SETI initiatives such as Breakthrough Lists.

Peter Xiangyuan Ma, an undergraduate researcher at DIAA and the University of Toronto’s Department of Mathematics and Physics, led the study. He was joined by researchers from UC Berkeley’s Radio Astronomy Laboratory, the Jodrell Bank Center for Astrophysics (JBCA), the Institute of Space Sciences and Astronomy, the International Center for Radio Astronomy Research, the SETI Institute, and Breakthrough Initiatives. The paper describing their findings, “A deep-learning search for technosignatures of 820 near stars,” recently appeared in Nature Astronomy.

Remove all ads on Universe today

Join our Patreon for just $3!

Get the ad-free experience for life

Narrowband radio signals remain the most popular and sought-after technosignature due to the good propagation of radio waves in interstellar space. For SETI researchers, the challenge has always been to distinguish between possible transmissions, terrestrial RFI and radio waves from cosmic sources. For their study, Ma and his colleagues applied a beta-convolutional variational autoencoder algorithm to 820 unique targets observed by the GBT over 480 hours of observations in the sky.

“There are many perturbations in many of our observations,” Ma said in a Dunlap Institute news release. “We need to distinguish the exciting radio signals in space from the uninteresting radio signals from Earth.” “There is a lot of interference in many of our observations,” Ma said in a Dunlap Institute news release. “We need to distinguish the exciting radio signals in space from the uninteresting radio signals from Earth.”

Ma began work on this algorithm while he was still in high school, which he hoped would speed up SETI by streamlining the search for technosignatures. According to Ma, the algorithm combines two subtypes of machine learning — supervised and unsupervised learning — which he calls “semi-unsupervised learning.” This approach involves using supervised techniques to guide and train the algorithm so that it can generalize (using unsupervised techniques) and find hidden patterns in the data more easily.

Since joining the Dunlap Institute, Ma and his colleagues have trained the algorithm on simulated signals to distinguish between potential signals that could be extraterrestrial in origin and human-caused interference. They also compared Ma’s algorithm to various machine learning applications, their accuracy and false positive rates, and used this information to create the finished product. “I didn’t tell my team until after the paper was published that this all started as a high school project that wasn’t really appreciated by my teachers,” Ma added.

Aerial image of the South African radio telescope MeerKAT, part of the Square Kilometer Array (SKA). Photo credit: SKA

When they applied this algorithm to the GBT data, they discovered eight new interesting radio signals from five stars that are 30 to 90 light-years from Earth. These signals were missed by previous analysis that did not rely on machine learning. But for the SETI team, these signals are noteworthy for two reasons. dr Steve Croft, Project Scientist for Breakthrough Listen on the GBT explained:

“First, they are present when we look at the star and absent when we look away – in contrast to local disturbances, which are generally always present. Second, the frequency of the signals changes over time so that they appear far away from the telescope. It’s a bit like walking down a gravel path and finding a rock stuck in the tread of your shoe that seems to fit perfectly.”

When dealing with data sets containing millions of signals, some signals may exhibit the two properties only by accident. For this reason, the researchers are not yet convinced that the observed signals are extraterrestrial transmissions, although they appear to be what the team would expect. For starters, the signals were not found when the team performed follow-up observations using the GBT of the same stars. Until further observations are made and the signals are reacquired, they remain signals of interest.

dr Cherry Ng, a research associate at the University of Toronto’s DIAA and co-author of the paper, has been working with Ma on this project since summer 2020. According to Ng, machine learning is very important in a field like SETI, which is becoming increasingly dominated by big data:

“By sifting through the data with each technique, we may be able to uncover exciting signals. I’m impressed by how well this approach has worked in the search for extraterrestrial intelligence. With the help of artificial intelligence, I’m optimistic that we’ll be able to better quantify the likelihood of the presence of extraterrestrial signals from other civilizations.”

A composite image of the future SKA telescopes, including the SKA-Mid and SKA-Low stations in Australia with the precursor MeerKAT telescope dishes in South Africa. Photo credit: SKAO

Looking ahead, Ma and his team hope to improve their new algorithm and apply it to other observatories and their datasets. These include radio telescopes such as the MeerKAT array in South Africa and the soon-to-be-completed Square Kilometer Array (SKA), which will combine the MeerKAT and Murchison Radioastronomy Observatory (MRO) in Australia into a single, powerful array. With these and other tools at their disposal, Ma says the team plans to greatly scale up their machine learning approach.

“With our new technique, combined with the next generation of telescopes, we hope that machine learning can take us from finding hundreds of stars to finding millions,” he said. “With our new technique, combined with the next generation of telescopes, we hope that machine learning can take us from finding hundreds of stars to finding millions,” he said.

This new approach represents a leading candidate for accelerating SETI and other transient research into the age of data-driven astronomy. Over time, SETI research could keep pace with exoplanet discovery and astrobiological studies, where each newly discovered exoplanet is rapidly can be examined for signs of biosignatures and technosignatures.

Further Reading: Dunlap Institute, Natural Astronomy

Like this:

How Loading…

Comments are closed.