- Статьи
- Science and technology
- The Code of the Universe: How AI searches for life outside the Solar System
The Code of the Universe: How AI searches for life outside the Solar System
The search for habitable planets outside the Solar system is the main task that astronomers around the world are already solving today. Several super-powerful telescopes are used to solve this problem, including NASA's TESS and Kepler missions. The problem is that such devices generate a huge amount of data every day, which a person cannot analyze quickly. And here neural networks come to the aid of a person. Izvestia looked into how AI helps to explore the universe.
How does AI discover new worlds?
Modern astronomy is the science of super—powerful telescopes and new technologies. Every night, modern survey programs such as the Legacy Survey of Time and Space at the Vera Rubin Observatory in Chile generate enormous amounts of information — up to 20 terabytes per observation session. A person can master a similar amount of data if they watch several million films in one day. It is physically impossible to process such arrays manually, and this is where artificial intelligence comes to the rescue.
The benefits of AI are particularly evident in the search for exoplanets, that is, those located outside the Solar system. They are important because they help to better understand the Earth. In addition, with each new discovery, the chance of discovering a planet suitable for life increases. Two NASA space telescopes are currently set up for this task — the Kepler mission is engaged in in-depth study of a small area of the sky, while the TESS mission, on the contrary, scans all available space.
Kepler Mission Spacecraft
Together, they identify thousands of potential planets, but these data require careful verification. It can take years for a person to do such a job, during which programs will offer millions of new options, and the process will become endless. But neural networks can come to the aid of astronomers. A striking example is the new RAVEN (Ranking and Validation of ExoplaNets) algorithm, developed specifically for the TESS mission.
The program examines the array of data provided by the telescope, searches for potential exoplanets there and immediately checks them, separating real discoveries from false signals that may arise due to stellar activity, extraneous "noise" from the equipment used and other factors. Moreover, it does this faster and more accurately than a human could do.
The TESS Space Telescope in outer space (picture)
The effect of the algorithm is impressive: with its help, scientists were able to confirm the discovery of 118 new exoplanets and identify more than 2,000 promising candidates for planets.
Why do we need information about exoplanets?
The study of new planets grouped around other stars, based on approximately the same principle as planets in the Solar System are arranged around the Sun, allows us to better understand the laws by which the universe is arranged.
For example, to clarify the principles by which planets are distributed around a star, to understand how these systems form and evolve and why objects in them can migrate or lose their atmosphere.
All this allows us to better understand the relationship between the Sun and the Earth and to clarify the forecast for the future of our planet.
The maximum task is to find other planets suitable for life. But even if this is not yet possible, each new sample brings astronomers closer to answering an important question: how does an Earth-like planet arise and what makes it habitable.
What does machine learning provide?
The discoveries that the RAVEN algorithm makes are important far beyond academic research. They don't just add new data to catalogs, they help us understand where in the universe conditions suitable for life may exist.
Anastasia Ivanova, a researcher at the Department of Physics of Planets and Small Bodies of the Solar System at the Space Research Institute of the Russian Academy of Sciences, notes that the mission of the TESS telescope, which scans the entire sky, generates enormous amounts of information. Their processing by classical methods would take years.
"Using machine learning algorithms allows you to quickly filter out noise and focus on really interesting events," she explains. — These methods are universal: although they are being developed in the field of computer science, their technologies are effectively used in astronomy and other areas where big data analysis is required.
Grigory Tsurikov, a researcher at the Institute of Astronomy of the Russian Academy of Sciences, PhD, emphasizes the practical significance of such discoveries. In his opinion, the search for potentially habitable worlds is the main task in the study of exoplanets.
— Every new object increases our chances of discovering life outside the Solar system. With the launch of modern telescopes, it becomes possible to study their atmospheres for the presence of molecules that are indicators of biological activity. The discovery of biomarkers will be humanity's next important step in deep space exploration," he emphasizes.
The history of the search for exoplanets clearly shows how our understanding of the universe is changing due to the development of analytical methods. For example, at the beginning of the research, it was believed that all exoplanets were "hot Jupiters", large gas giants close to their stars. However, in reality, such objects were simply the easiest to spot. With the development of tools and algorithms, it has become possible to find small planets in the habitable zone — it is on such objects orbiting calm stars that life is most likely to exist.
"In the short term, we are expanding our knowledge of planetary systems and their formation processes," adds Ivanova. "In the long term, the accumulated data will help us choose targets for interstellar missions and the search for life. Without a correct catalog of exoplanets, space expansion would be blind.
Can AI replace the astronomer in the study of the universe
According to the general opinion of astronomers, a lot depends on how to use AI. A researcher at the Russian Academy of Sciences notes that if the data from the TESS mission is analyzed, there is a risk of missing unique or non-standard systems.
— If you use only AI without additional analysis, you may lose important information. Using algorithms speeds up data processing, but machine results always require critical human thinking," explains Ivanova.
AI acts as an accelerator and filter, it takes over routine work, processes huge amounts of data, and identifies potentially interesting objects. And humans retain a key role in interpreting the results, analyzing in detail, and explaining the physics of the processes.
— The most important thing is that it is the researcher who sets a scientific task for which AI can be used. In the coming years, we can expect further development of autonomous algorithms capable of detecting increasingly complex and weak signals, as well as systems that integrate AI into real—time control of telescopes and satellites.
However, even in this case, the human role will be crucial. The formulation of hypotheses, the verification of unusual discoveries and the explanation of their significance will remain with the researcher.
In fact, the future of astronomy is seen as a symbiosis. AI frees scientists from routine data processing, allowing them to focus on creative and analytical tasks, while humans set the direction of research and ensure scientific rigor.
Переведено сервисом «Яндекс Переводчик»
