CIOInsider India Magazine

We can better understand the cosmos if we know where and how much matter there is. Galaxy clusters, which can include hundreds to thousands of galaxies as well as plasma, hot gas, and dark matter, are the largest objects in the universe. The gravitational pull of the cluster ties these components together. Understanding such galaxy clusters is crucial for determining the beginning and current evolution of the cosmos.

Artificial Intelligence in Calculating Mass
The method of calculating the mass of enormous clusters of galaxies has been improved thanks to artificial intelligence, say astronomers from the Institute for Advanced Study and the Flatiron Institute and its associates. The AI demonstrated that researchers may now obtain considerably more precise mass estimates than before by adding a straightforward element to an existing equation.

The total mass of a galaxy cluster is arguably the most important factor in defining its characteristics. Yet, estimating this amount is challenging because galaxies cannot be weighed by putting them on a scale. The fact that dark matter, which accounts for a sizable portion of a cluster’s mass, is unseen adds to the difficulty. Instead, they infer a cluster’s mass from other measurable characteristics.

How Did Researchers Previously Estimate Galaxy Cluster Masses?
Early in the 1970s, Rashid Sunyaev and Yakov B. Zel’dovich, who is currently a distinguished visiting professor at the Institute for Advanced Studies School of Natural Sciences, invented a new method for calculating galaxy cluster masses. Their approach is based on the idea that as the matter is compressed by gravity, its electrons push back. The interaction between the electrons and the light’s photons is changed by the pressure on the electrons. The interaction produces new photons as photons from the Big Bang’s afterglow strike the squeezed material. The characteristics of those photons rely on how tightly the material is compressed by gravity, which in turn depends on the mass of the galaxy cluster. Astrophysicists can determine the cluster’s mass by counting the photons.

Because the changes in the photon characteristics depend on the galaxy cluster, this integrated electron pressure is not a perfect substitute for mass. Wadekar and his colleagues hypothesized that a machine learning technology called “symbolic regression” might uncover a more effective strategy. To discover which equation best fits the data, the program essentially tries out numerous combinations of mathematical operators, such as addition and subtraction, with different variables.

Wadekar and his colleagues fed a state-of-the-art simulation of the universe, complete with numerous galaxy clusters, to their AI algorithm. Then, using their technique, CCA research colleague Miles Cranmer searched for and found any other variables that would enhance the mass estimates.

AI helps find novel parameter combinations that human analysts might miss. For instance, while it is simple for human analysts to spot two key elements in a dataset, AI is better able to sort through large amounts of data, frequently uncovering unanticipated contributing factors. The machine learning community is currently heavily focused on deep neural networks.

They are quite effective, but the scientists explain that one disadvantage is that they are essentially a blackbox. What occurs inside them is beyond our comprehension. When anything in physics produces good results, we want to know why. Symbolic regression is advantageous since it examines a given dataset and produces easy-to-understand mathematical expressions in the form of straightforward equations. It offers a model that is simple to understand.

By adding a single new term to the current equation, the researchers’ symbolic regression program provided them with a new equation that could forecast the mass of the galaxy cluster more accurately. Wadekar and his associates then discovered a physical explanation by working backward from this AI-generated equation. They discovered that the presence of supermassive black holes at the centers of galaxies and other regions of galaxy clusters where mass inferences are less accurate are correlated. Their revised equation reduced the significance of the intricate cores in the calculations, which enhanced mass estimations. The galaxy cluster resembles a doughnut in shape. The revised calculation eliminates the doughnut’s jelly, which can cause more significant inaccuracies, and instead concentrates on the doughy outskirts for more reliable mass inferences.

Using tens of thousands of simulated universes from the CCA’s CAMELS suite, the researchers tested the equation that was discovered by AI. In comparison to the currently employed equation, they discovered that the equation reduced the variability in galaxy cluster mass estimations by about 20 to 30 percent for big clusters.

How does the New Equation Help?
The new equation may offer observational astronomers a better understanding of the masses of the objects they detect in upcoming galaxy cluster surveys. In the near future, several surveys focused on galaxy clusters are anticipated. The Simons Observatory, the Stage 4 CMB experiment, and the eROSITA X-ray survey are a few examples. We can increase the scientific return from these surveys by using the new equations.

Wadekar says, “We think that symbolic regression is highly applicable to answering many astrophysical questions. In a lot of cases in astronomy, people make a linear fit between two parameters and ignore everything else. But nowadays, with these tools, you can go further. Symbolic regression and other artificial intelligence tools can help us go beyond existing two-parameter power laws in a variety of different ways, ranging from investigating small astrophysical systems like exoplanets to galaxy clusters, the biggest things in the universe.”