Some 540 million years ago, various life forms all of a sudden began to emerge from the muddy ocean floors of planet earth. The Cambrian Explosion is referred to as the Cambrian period, and these aquatic creatures are our early ancestors.
All the complex life on Earth evolved out of these underwater creatures. Scientists think that all it took was an ever-slight increase in ocean oxygen levels above a threshold.
We might be on the brink of a Cambrian Explosion of artificial intelligence (AI). A burst of incredibly capable AI programs such as Midjourney, DALL-E 2, as well as ChatGPT have highlighted the rapid progress of machine learning in the past couple of years.
In virtually all fields of science, AI is now used to aid researchers in performing routine classification tasks. It’s likewise helping our team of stereo astronomers broaden the search for extraterrestrial life and also the results have been promising so far.
Discovering alien signals with AI
As scientists search for evidence of intelligent life outside of the planet earth, we have created an AI program that outperforms traditional algorithms in signal detection tasks. Our computerized intelligence was taught to look for signals which couldn’t be generated by natural astrophysical processes by looking at information from radio telescopes.
When we sent an in the past studied dataset to our AI, it discovered eight signals of interest that the classic algorithm had missed. For clarity, these signals are most likely not from extraterrestrial intelligence and tend to be more likely rare instances of radio interference.
Nevertheless, our results – published today in Nature Astronomy – highlight how AI methods are sure to play a continual role in the search for extraterrestrial intelligence.
Not so intelligent
AI algorithms don’t understand or think. They’re great at pattern recognition and have been found useful for tasks including classification, although they’re not good at problem solving. They only do the things that they are trained to do.
And so despite the fact that the idea of an AI detecting extraterrestrial intelligence seems like the plot of an exciting science fiction novel, both terms are flawed: Attempts to find extraterrestrial intelligence can not find immediate evidence of intelligence, and AI programs are not intelligent.
Radio astronomers instead search for radio “technosignatures.” These imagined signals might imply the presence of technology, and by proxy, the existence of a modern society with the ability to make use of technology for interaction.
For our investigation, we developed an algorithm that uses AI methods to classify signals as both radio interference or a genuine techno signature candidate. And our algorithm works much a lot better than we anticipated.
What our AI algorithm does
Attempts to find technosignatures are compared to looking in a cosmic haystack for a needle. Huge amounts of data are created by radio telescopes and there’s great interference from a variety of sources, including satellites, phones and WiFi.
Preferably, search engines must be able to differentiate genuine technosignatures from false positives and do it fast. Our AI classifier fulfills these criteria.
It was actually developed by Peter Ma, a student at the University of Toronto and principal author of our paper. To be able to produce a set of training data, Peter inserted simulated signals into the actual data then utilized this dataset to train an AI algorithm known as an autoencoder. The autoencoder learned to identify the important features in the information as it processed the information.
These attributes had been inputted into a random forest classification algorithm in a next step. This particular classifier produces choice trees to determine if a signal is noteworthy or merely radio interference – basically sorting out the technosignature needles from the haystack.
As soon as we trained our AI algorithm, we supplied it with over 150 terabytes of information from the Green Bank Telescope in West Virginia (480 observing hours). It’s determined 20,515 signals of interest that we subsequently needed to manually verify. 8 signals of these possess the attributes of techno signatures and couldn’t be linked to radio interference.
Eight signals, no re-detections
To verify these signals, we headed back to the telescope to re-watch all eight signals of interest. In our follow up examinations, we had been not able to identify any of them once again.
We’ve been in similar situations before. In 2020, we found a signal that turned out to be damaging radio interference. Even though we will continue to monitor these eight new candidates, probably the most probable explanation is that they were uncommon manifestations of radio interference: not aliens.
The radio interference issue is unfortunately not going anywhere. However as new technologies arrive, we are going to be much better prepared to deal with it.
Narrowing the search
Just lately, our crew deployed an effective signal processor in South Africa on the MeerKAT telescope. The MeerKAT uses a strategy known as interferometry to assemble its 64 dishes to create one single telescope. This particular technique is better able to figure out exactly where in the sky a signal comes from, which will drastically reduce false positives from radio interference.
If astronomers observe a technosignature that can’t be described as interference, it would suggest humans are not the sole developers of technology in the Galaxy. It could be probably the most significant discoveries ever made.
Nevertheless, if we find nothing, it doesn’t necessarily mean we’re the only intelligent species on the planet. Non-detection could also mean we haven’t looked for the right signal style or our telescopes are not sensitive enough yet to detect faint transmissions from distant exoplanets.
Before a Cambrian Explosion of discoveries can be made, we might have to cross a sensitivity threshold. If we are truly alone, we need to rather reflect on the distinctive beauty as well as fragility of life on Earth.
This article is republished from The Conversation under a Creative Commons license. Read the original article.
