Faced with massive astronomical data, artificial intelligence can demonstrate its capabilities. The international team led by Ge Jian, a researcher at the Chinese Academy of Sciences Shanghai Astronomical Observatory, successfully "dug treasure" in the quasar spectral data released by the third phase of the International Sloan Digital Sky Survey Project through the deep learning method of artificial intelligence. Recently, the authoritative international astronomical journal "Monthly Report of the Royal Astronomical Society" published relevant research results. According to Ge Jian, the "neutral carbon absorbers" in cold gases and dust in the universe are important probes for studying galaxy formation and evolution. However, the signal of neutral carbon absorption lines is weak and extremely rare, and can only be found in massive quasar spectral data. Using traditional search methods is time-consuming and laborious, like searching for a needle in a haystack. The research team used artificial intelligence deep learning methods to design neural networks, generate a large number of simulation samples based on actual observations of neutral carbon absorption line features, train deep learning neural networks, and use these "trained" deep learning neural networks to search for "neutral carbon absorbers" in the data released by the third phase of the International Sloan Digital Survey project. Through this innovative method, the research team quickly discovered 107 cases of "neutral carbon absorbers" in cold gas clouds within early galaxies in the universe. This sample size is nearly twice the maximum sample size obtained previously, and more weaker signals have been detected than before. Having discovered so many "neutral carbon absorbers" of cold gases, the research team superimposed these spectra together, greatly improving the ability to detect the abundance of various metal elements and directly measuring the partial loss of metal abundance caused by dust adsorption. The research results indicate that as early as the age of the universe was only about 3 billion years (the current age of the universe is about 13.8 billion years), these early galaxies carrying "neutral carbon absorbers" probes had undergone rapid physical and chemical evolution, entering a physical and chemical evolution state between the Large Magellanic Dwarf Galaxy and the Milky Way, producing a large amount of metals. At the same time, some metals were adsorbed onto dust, resulting in the observed "dust reddening" effect. "Our discovery independently confirms the new discovery of diamond like carbon dust detected by the James Webb Space Telescope in the earliest stars in the universe, indicating that some galaxies have evolved much faster than expected, which will challenge the existing models of galaxy formation and evolution," said Ge Jian. Industry experts believe that this research is an important breakthrough in the application of artificial intelligence in the field of astronomical big data. Artificial intelligence deep learning methods have enormous application value and potential in multi domain image recognition and weak signal detection. In the future, it is expected to mine more treasures from massive astronomical data. (Lai Xin She)