The Gaia telescope works like a cosmic sieve. It measures photometric data (brightness in different colors) and spectra (the rainbow breakdown of light). A neural network is trained to distinguish patterns: which ones belong to stars, and which to galaxies. For bright objects, accuracy exceeds 96%.

But faint distant sources, like grains of sand, slip through the sieve. That's where the infrared telescope CatWISE2020 comes in — it sees thermal radiation that penetrates even dust clouds. It acts as a finer-mesh sieve, allowing capture of nearly a third more hidden quasars and galaxies. As a result, the Gaia DR4 catalog gained three million quasar candidates and two million galaxies, though among faint sources there may be some false positives.

This cosmic census helps us understand how the Universe evolves in its farthest reaches.