UC Berkeley scientists focus on 100 radio signals in search of extraterrestrial life
FILE ART- (Photo by Planet Observer/Universal Images Group via Getty Images)
BERKELEY, Calif. - What does it take to detect a radio signal sent by extraterrestrial life to Earth? Two decades of work involving radio telescopes stationed on opposite sides of the world, a supercomputer in Germany, and two million volunteer computers worldwide.
Maybe.
A team of UC Berkeley scientists has painstakingly analyzed billions of radio signals received over fifteen years and have zeroed in on 100 that are most likely to be sent by extraterrestrial intelligence.
The project -- called SETI@home after the Search for Extra-Terrestrial Intelligence -- sought to scan a third of the sky visible from Earth to isolate a certain type of radio signals: those belonging to a narrow bandwidth, with most of their power concentrated at a small range of frequencies. Scientists believe these characteristics could indicate that they are sent by an extraterrestrial life form.
It took two decades for the project to arrive at its conclusion, but as a result, a dataset of 12 billion radio signals has been winnowed down to the 100 that are likeliest to have been sent by an extraterrestrial life form.
The results of this project were published in two papers in The Astronomical Journal in 2025.
SETI@home was conceived in 1995, when David Anderson, a computational scientist and computer science professor at UC Berkeley was researching "distributed computing." This technique entails distributing large computational tasks into smaller tasks across a large network of computers, significantly cutting down the costs for researchers.
"1995 was a time when home computers were getting connected to the internet," said Anderson, "and it was also a time when home computers were getting pretty fast."
Pilots report UFO sightings during flight | West Coast Wrap
Tonight -- a strange sight in the skies over Oregon--pilots report seeing several UFOs. Coming up -- hear how they describe those objects in calls to the control tower. Plus -- a California couple killed while on vacation in Mexico. What we're learning about the shooting -- as investigators search for a motive.
When a graduate student suggested using distributed computing to analyze radio signals received by radio telescopes to scan for extraterrestrial intelligence, Anderson teamed up with UC Berkeley astronomers Erik Korpela and Dan Werthimer to work on the project.
The intention was to look for extraterrestrial life outside the solar system that might be sending radio signals to Earth to draw attention to themselves.
First, the three scientists found a suitable SETI research project at UC Berkeley that had access to data from the radio telescope at the Arecibo Observatory in Puerto Rico.
Then they developed the SETI@home software, which volunteers worldwide could download on their own computers.
Like a screensaver, this software was designed to work in the background. It would analyze the radio signals from Arecibo and identify those that adhered to certain parameters that scientists think are consistent with a signal that intelligent extraterrestrial life might send.
The software was launched in 1999, and the response was unexpected -- within a year, 2 million people had downloaded the software. Anderson said that they only started receiving good quality data in 2006, after an upgrade to the Arecibo telescope.
Between 2006 and 2020, the telescope scanned a third of the sky visible from Earth multiple times, supplying the home computers with enough data to compile a list of 12 billion signals that fit the criteria that the team was looking for.
Expecting this avalanche of data and the need to filter it further, in 2016, Anderson, Korpela, and Werthimer started working on specialized algorithms to eliminate the spurious signals generated by interference from Earth's radio, TV stations, and aircraft radar systems.
For that, they used the supercomputer at the Max Planck Institute for Gravitational Physics in Hanover, Germany. Running the algorithms on the supercomputer only took two days, said Anderson, but the whole process took two whole years.
"We had to do that hundreds of times, because... it was all very experimental," he said. "Nobody had ever done anything like this before, so we started off with some algorithms that did not work very well. It was all trial and error."
Once this second round of sorting and ranking was complete, the first being done by home computers, Korpela and Wertheimer examined the top thousand candidates manually, and further thinned the list to the top 100 viable signals.
Now, the sources of these 100 signals are being observed through the Five-hundred-meter Aperture Spherical Telescope (FAST) in China -- 15 minutes for each point of origin -- to test if the signals are repeated.
Anderson said that so far, the sources of half of the 100 signals have been observed, and they haven't returned a hit. Chances of the remaining half yielding a hit are slim, but he believes that SETI@home has been groundbreaking in the scope of the signals it analyzed and the precise algorithms it produced to eliminate radio interference.
Above all, he sees the project as a successful demonstration of how distributed computing can democratize science.
"The [computing] power of SETI@home would cost hundreds of millions of dollars if you bought it on the Amazon cloud. Volunteer computing makes it possible to do that kind of science for a lot less money," he said.
"My personal hope is that... companies like OpenAI and Google that are able to spend a billion dollars to build new data centers will not have a monopoly on AI if we can figure out a way to use volunteer computing to train large language models. That could be a game changer for AI!"