Artificial intelligence is an ideal tool for exploring the Universe

In an attempt to understand the universe, we become obsessed - we are drawn to the thirst for control. Satellites transmit hundreds of terabytes of this information each year, and only one telescope in Chile will produce 15 terabytes of satellite imagery every night. No one can handle them manually. As astronomer Carlo Enrico Petrelo says, "Looking at galaxies is the most romantic part of our work." The problem is how to maintain focus. Therefore, he develops artificial intelligence, which will help him.

Pittrello and his colleagues were looking for a phenomenon that is essentially a space telescope. When a huge object (galaxy or black hole) is between a distant light source and an observer on Earth, it bends space and light around it, creating a lens that allows astronomers to take a closer look at the incredibly old and distant parts of the universe hidden from our view. This effect is called gravitational lenses, and these lenses are the key to understanding what the universe is made of. So far, it was slow and tedious to look for.

This is where artificial intelligence is needed - and the search for gravity lenses is the beginning. As Stanford University professor Andrew Eun said, Amnesty International's ability to automate everything "can be done by an ordinary person in less than a second of thinking." It may not seem less than a second particularly generous, but when it comes to posting large amounts of data, it is just a heavenly gift.

The new wave of astronomers is regarded as artificial intelligence not only as data capture. They explore something that could be a whole new way to find scientific discoveries, when artificial intelligence reflects parts of the universe that we have never seen before.

But first: gravity lenses. The general theory of Einstein's relativity predicted this phenomenon in the 1930s, but the first examples only emerged in 1979. Why? Because the space is very large, people need a lot of time to look around, especially without the use of modern telescopes. The search for gravity lenses was complex.

"The lenses we have now have been found in many ways," says Lilly Williams, a professor of astrophysics at the University of Minnesota. "Some were discovered by chance, and people were looking for something completely different ... Some were found by people who were looking for them, second or third."

See the pictures Ai can do a good job. Thus, Petrillo and his colleagues turned to the artificial intelligence tool, beloved in the Silicon Valley: a type of computer program consisting of digital "neurons" that have been modeled after the present and activated in response to input. Feed these programs (neural networks) a lot of data - they will learn how to recognize patterns and patterns. It works well with visual information and is used in a variety of automated vision systems - from auto-controlled cameras to facial recognition in Facebook images.

Our New YouTube Video

Please subscribe to our youtube channel BHASKBAN

Video Embed by :Bhaskban

As he wrote in an article published last month, the application of this technique to hunting for gravitational lenses was surprisingly simple. First, the scientists prepared a set of data to train the neural network - generated 6 million fake images with and without attractive lenses. The neural network then feeds their data and leaves to understand the patterns. Fine tuning and a program that recognizes gravity lenses in the blink of an eye.

"An excellent work is done in the person of the photo man at a rate of one thousand per hour," says Petrillo. There is one lens once in 30,000 galaxies. Therefore, the workbook will have to work without sleep and rest for a week to find five or six lenses only. The neural network, for comparison, understands 21,789 images in just 20 minutes. This is with one old processor.

The neural network was not as precise as the computer. To avoid looking at the lens, the criteria were given wide. It issued 761 potential candidates, which people studied and reduced to 56. In order to ensure that these lenses are real, they will have to verify and verify the results, but Petrillo believes that one-third will look real. It turns out that there is one lens per minute, if compared with 100 lenses, discovered by the entire scientific community over the past few decades. The speed is incredible and the prospects are huge.

The search for these lenses is necessary to understand one of the greatest mysteries of astronomy: What is the universe? The material we know (planets, stars, asteroids, etc.) represents only 5% of the total material, and 95% is completely unreachable. These 95% are represented by a hypothetical substance - the dark matter, which we have not observed directly. We can only study the effect of gravity on the rest of the universe, and gravity lenses are one of the most important indicators.

What can Amnesty International do? Scientists are working on a number of new tools. Some, like Petrillo, carry out the task of identification: they classify galaxies, for example. Others scan data streams for interesting signals. Some neural networks require artificial interference in the radio telescope, isolating only useful signals. Others used to identify pulsars, unusual external planets or improve low-resolution telescopes. In short, there are many potential applications.

This explosion is partly due to general trends in hardware, which allow the expansion of artificial intelligence, such as the availability of cheap computing power. Astronomers no longer need to sit on their pants with net s, watching the movements of individual planets. Instead, they use advanced technology that scans the sections of the sky one by one. Williams says improved telescopes and data storage technologies mean the potential for analysis is greater.

Analysis of large data sets - this is what artificial intelligence can do. We can teach him to recognize patterns and make him work tirelessly, never closing or making a mistake.

Do astronomers worry that they trust a machine that may lack the human understanding to discover something exciting? Petrillo says that no. "In general, people are more biased, less effective and more prone to error than machines." Deliver agreed. "Computers can miss certain things, but they will miss them on a regular basis." But as long as we know what they do not know, we can deploy automated systems without much risk.

For some astronomers, the potential of artificial intelligence goes beyond simple data counting. They believe artificial intelligence can be used to create information that fills blind spots in our observations about the universe.

Astronomer Kevin Shavensky and his team of astrophysicists in galaxies and black holes used artificial intelligence to improve the accuracy of dim telescope images. To this end, they have published a neural network that unambiguously generates variations in the data under study, as if a good fake imitates the style of a famous artist. The same networks were used to create fake images of stellar images; mock phonetic dialogues that mimicked real sounds; and other types of data. According to Shavensky, such neural networks create information that was not previously accessible.

In a paper published by Shavinsky and his team earlier this year, they showed that these networks can improve the quality of space images. They reduced the image quality of a number of galaxies, noise and noise, and then passed it along the neural network along with the original images. The result was amazing. But scientists can not share them yet.

Shavinsky warns of the project. After all, it conflicts with the basic principles of science: you can recognize the universe only by observing it directly. "For this reason, this tool is dangerous," he says. It can only be used when we have accurate data and when we can verify the result. You can train a neural network to generate data about black holes and send them to work in a certain section of the sky, which has been researched so far poorly. If there is a black hole, astronomers will have to confirm the result with their hands - as in the case of gravitational lenses.

If these methods prove fruitful, they can become completely new research methods, complementing classic computer modeling and good old observation. While everything has just begun, but the possibilities are very promising. "If you have this tool, you can take all the data from the archive, improve some of them and get great scientific value." Value that was not before. Amnesty International will become a scientific chemical that will help us transform old knowledge into new knowledge. We can explore the universe as never before, without leaving Earth.

No comments

Powered by Blogger.