Domestic Astronomical AI Model Breaks the Limit of Observation Depth

Before the vast and boundless universe, human vision has always had a "ceiling." Even with powerful instruments like the James Webb Space Telescope, it often struggles to detect faint signals from the early universe due to background noise. Recently, a research breakthrough from Tsinghua University has broken this impasse. Developed by the team led by Academician Dai Qionghai from the Department of Automation and Associate Professor Cai Zheng from the Department of Astronomy, the AI astronomical observation enhancement model called "Xingyan" successfully solved the problem of high-fidelity photon reconstruction under extremely low signal-to-noise ratios, achieving a significant breakthrough in astronomical observation depth.

How strong is this domestic "black technology"? Experimental data shows that the "Xingyan" model directly improved the detection depth of the James Webb Space Telescope by 1 magnitude, and the detection accuracy increased by 1.6 magnitudes. In astronomy, these seemingly small numbers mean we can capture more distant and fainter starlight. With this tool, the research team discovered over 160 candidate high-redshift galaxies from the early universe in the deep observation data of the James Webb, three times more than previous discoveries.

These newly discovered galaxies were born during the "cosmic dawn," just 200 million to 500 million years after the Big Bang. They are like the first glimmers of light in the universe's infancy, providing precious original data for understanding the origin of all things.

From a technical perspective, the biggest enemy of astronomical observation is noise. Bright sky backgrounds and the telescope's own thermal radiation act like a thick fog, obscuring faint starlight. The magic of "Xingyan" lies in its ability to reconstruct a flat deep-space image into a three-dimensional volume intertwined with time and space, and through a unique luminosity-adaptive filtering mechanism, it precisely separates noise from target signals. This "clearing the fog to see the blue sky" capability not only accurately restores signals but also redefines the boundaries of deep-space exploration.

Currently, the related research results have been published in the prestigious international academic journal "Science." This is not only a victory of the deep integration of AI technology and basic science, but also a solid footprint left by Chinese research teams in their journey to explore the stars and oceans. With the further application of "Xingyan," the secrets of the deepest parts of the universe may be coming to us faster than ever before.