Chinese scientists push the limits of deep space exploration with new AI model

Chinese scientists have developed an advanced astronomical artificial intelligence (AI) model that has empowered them to capture the the deepest views of space.

Named ASTERIS, the model employs cutting-edge computational optics and AI algorithms to detect faint celestial signals, allowing astronomers to observe galaxies located more than 13 billion light-years away. A light-year is about 9.46 trillion kilometers.

The breakthrough was published in the journal Science on Friday.

Faint celestial bodies hold critical clues about the origin and evolution of the universe. However, studying these objects has been a challenge due to interference from background sky noise and thermal radiation from telescopes.

A cross-disciplinary team from Tsinghua University developed the AI model to address the challenges. The model is capable of decoding vast amounts of data from space telescopes and is compatible with multiple detection devices, offering a potential universal platform for enhancing deep space data analysis.

The team demonstrated that when applied to the James Webb Space Telescope, the model extended the telescope's observation range from visible light at around 500 nanometers to the mid-infrared at 5 micrometers. It also increased the detection depth by 1.0 magnitude, effectively enabling the telescope to detect objects 2.5 times fainter than previously possible, which is like increasing the size of the telescope's aperture from approximately 6 meters to nearly 10 meters.

Using the model, the team identified over 160 candidate high-redshift galaxies from the "Cosmic Dawn" period, existing between 200 million and 500 million years after the Big Bang – a major leap from the previously known some 50 galaxies from that period, according to Cai Zheng, associate professor at Tsinghua's Department of Astronomy and a member of the research team.

Traditional noise-reduction techniques rely on stacking multiple exposures and assume noise is uniform or correlated. In reality, deep-space noise varies across both time and space. ASTERIS addresses this issue by reconstructing deep-space images as a 3D spatiotemporal volume.

Through "photometric adaptive screening mechanism," the model identifies subtle noise fluctuations and distinguishes them from the ultra-faint signals of distant stars and galaxies.

"Overall, I think this is a very relevant piece of work that can have an important impact across astronomy," one reviewer of the research said.

The model allows for the high-fidelity reconstruction of faint celestial bodies affected by light noise, said Dai Qionghai, professor at Tsinghua's Department of Automation. This breakthrough technology is expected to be applied to next-generation telescopes, supporting the exploration of dark energy, dark matter, the origin of the universe, and exoplanets.

(With input from Xinhua)