Chinese scientists have developed a new AI model that allows researchers to analyze stellar data collected by different telescopes in a unified way, easing one of astronomy’s long-standing data challenges. The AI model enables spectral information from different telescopes to be compared and studied together, improving large-scale research on the Milky Way.
The model, named SpecCLIP, was created by researchers from the National Astronomical Observatories of the Chinese Academy of Sciences, the University of Chinese Academy of Sciences, and partner institutions. The work has been reported by Science and Technology Daily and published in the Astrophysical Journal.
Astronomers rely on stellar spectra to understand stars. These spectra reveal key physical details, including temperature, chemical makeup, and surface gravity. By studying such data across vast numbers of stars, researchers can reconstruct how the Milky Way formed and evolved over billions of years.
AI model links stellar data from different telescopes
Until now, combining spectral data from different surveys has been difficult. Projects such as China’s LAMOST telescope and Europe’s Gaia satellite collect data using different methods, resolutions, and wavelength ranges. This makes direct comparison unreliable and limits joint analysis across missions.
Chinese astronomers just dropped SpecCLIP, an AI “translator” that unifies stellar spectra from LAMOST (low-res) and Gaia (high-precision)! It turns mismatched data into a universal language for Galactic archaeology, hunting metal-poor ancient stars, and even exoplanet screening. pic.twitter.com/GIgxH8aYW2
— Tom Marvolo Riddle (@tom_riddle2025) February 25, 2026
SpecCLIP was designed to remove that barrier. The research team applied concepts similar to those used in large language models and trained the system through contrastive learning. This allows the AI to identify and learn the underlying relationships between spectra from different sources without manual alignment.
Huang Yang, an associate professor at the University of Chinese Academy of Sciences and corresponding author of the study, explained that the model works like a translator.
It converts low-resolution spectra from LAMOST and high-precision spectra from Gaia into a shared representation, allowing scientists to align and analyze the data across instruments and survey programs.
Broad capabilities and early scientific use
Unlike traditional tools built for a single task, SpecCLIP functions as a broader framework. It can predict stellar atmospheric parameters and elemental abundances in one step, search for stars with similar spectral features, and assist in identifying unusual celestial objects.
These capabilities are especially important for Galactic archaeology. Researchers aim to find extremely rare, metal-poor ancient stars that preserve clues about the Milky Way’s earliest stages. SpecCLIP can help sift through massive datasets to locate such objects more efficiently.
The AI model has already been used in advanced research projects. In the Earth 2.0 mission, which searches for Earth-like planets, SpecCLIP helps accurately characterize stars that host planets, improving the screening process for potentially habitable worlds.
Researchers say the project highlights the growing role of AI in astronomy and its ability to integrate massive datasets from different telescopes, opening the door to more precise studies of the galaxy’s structure and history.
