March 28, 2023

Artificial Intelligence detects strange signals in radio data

In recent years, an explosion of incredibly capable Artificial Intelligence (AI) programs such as Midjourney, DALL-E 2, and ChatGPT have demonstrated the rapid progress the science of machine learning has made.

The artificial intelligence algorithm developed by a team of scientists at Curtin University in Australia has detected strange radio signals in the universe. Image: arleksey – Shutterstock

Artificial intelligence is now used in virtually every field of research to help scientists with routine classification tasks.

publicity

According to astrophysicist Danny Price, in an article he wrote and published on the site The conversationartificial intelligence is also helping radio astronomers expand the search for extraterrestrial life, and the results so far have been promising.

“As scientists looking for evidence of intelligent life beyond Earth, we have built an artificial intelligence system that outperforms classical algorithms in signal detection tasks,” said Price, who is a senior researcher at Curtin University in Australia. “Our AI was trained to search through radio telescope data for signals that could not be created by natural astrophysical processes.”

He says that when his team fed the AI ​​a previously studied data set, the system discovered eight signals of interest that the classic algorithm missed. “To be clear, these signals are probably not alien intelligence and are more likely rare cases of radio interference.”

However, these findings – were published on Monday (30) in the journal Astronomy of Nature — “point out how artificial intelligence techniques will certainly play a lasting role in the search for extraterrestrial intelligence,” Price said.

AI is not that smart

Price explains that AI algorithms do not “understand” or “think.” They excel at pattern recognition and have proven extremely useful for tasks such as classification – but lack the ability to solve problems. “They’re just doing the specific tasks they were trained to do.”

So, although the idea of ​​an AI detecting alien intelligence sounds like the plot of an exciting science fiction movie, to the astrophysicist both terms are wrong: “AI programs are not intelligent, and searches for alien intelligence cannot find direct evidence of intelligence’. .

Instead, radio astronomers look for radio “artifacts.” These hypothetical signs would indicate the presence of technology and the existence of a society with the ability to exploit technology for communication.

Read more:

“For our research, we created an algorithm that uses artificial intelligence methods to classify signals as either radio interference or a genuine technology signature candidate,” Price said. “And our algorithm performs better than we expected.”

What the AI ​​algorithm developed by Price does

According to the researcher, radio telescopes produce huge amounts of data and there is huge interference from sources like phones, WiFi and satellites. Search algorithms must be able to filter out real technology signatures from “false positives” – and do it quickly.

“Our AI classifier meets these requirements. It was designed by Peter Ma, a student at the University of Toronto and lead author of our article. To create a training dataset, Peter fed simulated signals into real data and then used that dataset to train an AI algorithm called an autocoder.

As the autoencoder processed the data, according to Price, it “learned” to recognize features that appeared in the data.

In a second step, these features were fed into an algorithm called a random forest classifier. “This classifier builds decision trees to decide whether a signal is noticeable or just radio interference.”

The Green Bank Telescope in the US detected more than 20,500 radio signals with the help of the Artificial Intelligence algorithm. Image: John M. Chase – Shutterstock

After training the AI ​​algorithm, Price’s team fed it more than 150 terabytes of data (480 hours of observation) from the Green Bank Telescope in West Virginia. 20,515 signals of interest were identified, which had to be checked manually. “Of these, eight signals had the characteristics of technological signatures and could not be attributed to radio interference,” the researcher said.

“We have found similar situations again. In 2020, we detected a signal that turned out to be harmful radio interference. While we are tracking these eight new candidates, the most likely explanation is that they were unusual radio interference events: not aliens.”

Watch our new videos on YouTube? Subscribe to our channel!

Leave a Reply

Your email address will not be published. Required fields are marked *