The international team led by Ge Jian, a researcher at the Chinese Academy of Sciences Shanghai Astronomical Observatory, used the deep learning method of artificial intelligence to conduct weak signal search and data analysis on the quasar spectral data released during the third phase of the International Sloan Sky Survey, and found the key probe neutral carbon absorber of the cold gas cloud composition in 107 rare early galaxies in the universe. Analysis has found that in the early stages of the universe's evolution of about 3 billion years, these early galaxies carrying neutral carbon absorbing probes have undergone rapid physical and chemical evolution and entered a physical and chemical evolution state between the Large Magellanic Dwarf Galaxy and the Milky Way. This study provides a new research approach for exploring how galaxies form and evolve, demonstrating the application prospects of artificial intelligence in exploring weak signals in massive astronomical data.

Studying cold gases and dust is crucial for analyzing galaxy formation and evolution, providing astronomers with a crucial means to explore the drastic changes in galaxies from their initial assembly to star formation, and then to the entire lifecycle of galaxies in the later stages of evolution. Neutral carbon absorption lines can serve as effective probes to track gas and dust components. And the signals of these neutral carbon absorption lines are weak and rare, which can only be found in massive quasar spectral data. The use of traditional search methods is time-consuming, and at the same time, more false signals are detected, making it easy to miss some weak signals.

The team uses 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, and train deep learning neural networks. These "trained" deep learning neural networks are used to search for neutral carbon absorbers in the data released during the Sloan Survey Phase III. The team discovered the key probe for the composition of cold gas clouds in 107 rare early universe galaxies through the above method - neutral carbon absorbers. The sample size obtained in this study is nearly twice the maximum sample size previously obtained, and more weaker signals were detected than before.

This study found neutral carbon absorbers with more cold gases and overlaid these spectra to directly measure the partial loss of metal abundance caused by dust adsorption. The results indicate that when the universe was only about 3 billion years old, these early galaxies carrying neutral carbon absorbing 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. At this point, a large amount of metal is produced, and some of the metal is adsorbed onto the dust, resulting in observable dust reddening. This confirms the recent discovery of diamond like carbon dust by the James Webb Space Telescope in the earliest stars in the universe, indicating that some galaxies have evolved much faster than expected, challenging existing models of galaxy formation and evolution. Unlike the James Webb Space Telescope, which conducts research through galaxy emission spectra, this work investigates early galaxies by observing the absorption spectra of quasars. This will provide a new research tool for future research on the early evolution of the universe and galaxies.

Ge Jian said, "To use artificial intelligence to 'dig' new discoveries in massive astronomical data, it is necessary to develop innovative artificial intelligence algorithms that can quickly, accurately, and comprehensively explore these rare and weak signals that are difficult to find in traditional methods."


On May 15th, the relevant research results were published in the Monthly Report of the Royal Astronomical Society (MNRAS).



Paper link