This system searches for aliens using AI: it detected 8 “very suspicious” radio signals

This system searches for aliens using AI: it detected 8 “very suspicious” radio signals

From a sample of more than 20,000 signals, the system was able to pick up 8 candidates for artificial radio signals that attempts were made to investigate years ago.

system searches for aliens using AI

Artificial intelligence has awakened thousands of applications hitherto unthinkable, also in Spain. With the appearance of ChatGPT and Microsoft’s sudden interest in this technology, Google has seen its foundations shake to the point of causing a war between the two companies for the inclusion of AI in their ecosystems. But artificial intelligence could go much further, as a new study proposes its use to help identify extraterrestrial radio signals.

Peter Ma, a student at the University of Toronto in the United States, has developed a new machine-learning algorithm to filter out radio signals from outer space to identify those likely to have alien origins. The most curious thing is that Ma has managed to discover “eight signals that seemed very suspicious,” according to Nvidia’s blog.

The study, published in the journal Nature Astronomy, highlights the role that artificial intelligence technologies will play in the search for extraterrestrial life and intelligence, which could lead us to a hypothetical future in which we can find artificial space radio signals.

PostGoogle’s answer to ChatGPT that will revolutionize your search engine

AI to search for aliens

We have what radio signals mean it is easy to think that these are like those that humans emit through radio frequency signals due to the idea. However, the universe is by no means a silent place. Numerous natural astrophysical events, such as stars, black holes, magnetars, or supernovae, generate radio and many other signals.

Added to this is the interference suffered by our technology, which makes it difficult to detect genuinely artificial radio signals. And the possibility of Earth-like life is incredibly slim. Both Ma and his idea are encompassed in Breakthrough Listen, an international effort looking for signs of extraterrestrial civilizations. In this case, find genuinely artificial signals.

Space photography

The system is in charge of filtering the most promising technological candidates. It is produced in a two-step process. The first uses an automatic encoder to identify salient features in the data. A system powered by the TensorFlow API and accelerated by four Nvidia TITAN X graphics cards at UC Berkeley. The second step feeds those features into a system called a “random forest classifier,” which decides whether the signal is noticeable or just interference.

Once those signals are collected, the researchers return to the telescope to look at the systems from which the eight signals originated. Signals that have not been able to be observed again in at least several years. The system was fed with a massive data set, more than 150 terabytes of data collected from the Green Bank radio telescope, one of the largest in the world, in West Virginia.

PostGoogle Bard: the Artificial Intelligence that promises to leave ChatGPT behind is official

“Suspicious Signals”

The “eight signals looked very suspicious after taking another look at the targets with our telescopes. It’s been almost five or six years since we took the data, but we still haven’t seen the signal again. 

These 8 signals were filtered out of more than 20,000 signals of interest, showing telltale features that scientists call ‘technological signatures’, which could not be interference. The system developers themselves admit that because they trained on simulated signals, it is likely that the system can learn from signals that do not exist. One of the paper’s co-authors, Cherry NG, says the system is capable “of generating signals that aliens may or may not have sent.”

Both Ma and Cherry hope that study will serve to reveal the potential of AI in the investigation of signs of extraterrestrial life and the importance of analyzing large amounts of data. Ma said, “We hope to extend this search capability and algorithm to other types of telescope configurations”.

Related PostMicrosoft wants to dethrone Google with its new vitaminized search engine with ChatGPT

Leave a Reply

Your email address will not be published.