Imagine a painting made of millions of tiny dots in different colors. From a distance, the dots blend into a single tone, but by the shades in different areas, you can guess which colors dominate. That's how astronomers work: they look at the collective light of distant galaxies and, from its color, figure out how these galaxies are distributed by distance.

The universe is full of galaxies, but most are too dim to see individually even with the most powerful telescopes like Hubble and James Webb. Their combined light merges into a faint glow — the cosmic optical background. If you look closely at this background in different colors (called photometry), you can spot a subtle ripple. It arises because distant galaxies shine slightly differently due to the expansion of the universe: their light gets stretched, turning redder. It's like the sound of an ambulance speeding away: the farther it is, the lower the pitch. In astronomy, this shift is called redshift, and it helps gauge an object's distance.

The new method uses machine learning — a computer program that learns from examples. First, scientists created a vast virtual universe through numerical simulation, where all distances to galaxies were known in advance. Then they showed the program hundreds of thousands of background images along with the correct answers. The program learned to guess how far away those clumps of light are. As a result, the accuracy proved better than one percent — even for galaxies that can't be seen individually.

This discovery will help us better understand dark matter — the invisible mass first proposed after the observations of Vera Rubin. It will also let us study the era when the first stars ignited, the epoch of star formation, and how interstellar dust dimmed their light. The idea traces back to the discovery by Edwin Hubble, who proved the universe is expanding, and the dreams of Nancy Grace Roman about powerful space telescopes. By the way, this method is reminiscent of studying the cosmic microwave background, but instead of ancient microwave light, we're analyzing visible light.