The universe is estimated to contain hundreds of billions of galaxies, and each galaxy has roughly the same number of planets, so the chances of Earth being the only life form are extremely small. Scouring millions of radio signals from space to identify those with potential man-made origins, they discovered eight intriguingly alien-looking signals.
When aliens scan our planet with the right equipment, they receive radio waves and other electromagnetic signals that we have been transmitting for the better part of a century. With that in mind, the Breakthrough Listen initiative aims to turn the tables and search for artificial radio signals coming from other planets in the galaxy. The team calls these signals “techno signatures”.
The problem is that space is a noisy place. Stars, black holes, magnetars, quasars, FRBs, supernovae, gamma ray bursts, and various other objects and events can produce radio waves and other signals. In addition, there is interference from our own technologies, such as mobile phones and GPS satellites. Adjusting for background noise to find possible alien techno signatures is a monumental task.
Or at least, it’s for humans. Artificial intelligence is adept at sorting through vast amounts of data to look for patterns. So this is the perfect job to do it. For the new study, Peter Ma, a student at the University of Toronto, developed a new machine learning algorithm designed to weed out the most promising technical signature candidates.
This algorithm works in a two-step process. The first step involves an autoencoder. Autoencoders are trained on intelligent, simulated extraterrestrial signals so they know what to look for. Essentially, it should be a narrowband signal with a detectable drift rate, visible only in observations of certain regions of the sky. These simulated signals are added to a pool of real data until the autoencoder can reliably select them.
Once it can do that, the AI is freed up for real work. Each signal in the dataset is run through an algorithm called a random forest classifier to sort the interesting from the noise. In this case, the team fed the AI system with over 150 TB of data collected by the Green Bank Telescope (GBT).
Of the 3 million signals in the dataset, the AI identified 20,515 signals of interest. Researchers then had to manually inspect each of these signals. Interestingly, eight of these signals had good characteristics for technical characterization and were not due to interference.
“The eight signals looked very suspicious, but after another look at the target with the telescope, we never saw them again.” It’s been almost five to six years since we got the data. but the signal is not yet confirmed, make it what you want.
The signal is interesting, but it’s nowhere near the answer to the most serious question of whether we are alone in space. Admittedly, using simulations can train the AI to focus on the wrong things.
Still, it is a worthwhile exercise, and applying AI to other datasets may yield more possible technical features.
A study was published in a journal natural astronomy.
Source: NVIDIA, Breakthrough Listen
