中国AI 模型SpecCLIP将望远镜上的恒星数据标准化,帮助发现古代恒星和类似于地球的行星。 Chinese AI model SpecCLIP standardizes stellar data across telescopes, aiding in discovering ancient stars and Earth-like planets.

中国研究人员开发了名为SpecCLIP的AI模型,将不同望远镜的恒星光谱数据转换成共同格式,克服LAMOST和Gaia等调查数据不一致的问题。 Chinese researchers have developed an AI model called SpecCLIP that translates stellar spectral data from different telescopes into a shared format, overcoming inconsistencies in data from surveys like LAMOST and Gaia. SpecCLIP利用大语言模型概念和对比性学习,能够对庞大的天文数据集进行无缝分析,预测恒星特性,识别稀有的古星,并支持搜索类似于地球的行星。 Using large language model concepts and contrastive learning, SpecCLIP enables seamless analysis of vast astronomical datasets, predicting stellar properties, identifying rare ancient stars, and supporting searches for Earth-like planets. 该模型在《天体物理杂志》上发表,作为银河考古学的基础工具,并突出大赦国际在推进大规模天文学发现方面发挥的日益扩大的作用。 The model, published in the Astrophysical Journal, serves as a foundational tool for Galactic archaeology and highlights AI’s expanding role in advancing large-scale astronomical discovery.