Science News内容：

　　题 目：AstroLLaMA : Towards Specialized Foundation Models in Astronomy

　　摘 要：AstroLLaMA is a 7 billion-parameter language model fine-tuned on a dataset encompassing over 300,000 abstracts from astronomical research papers. Compared to GPT-4, a current state-of-the-art general Large Language Model, it shows superior context- aware capabilities of astronomical concept despite having much fewer parameters.

　　报告人：刘鹏辉（紫台研究生）

Journal club内容：

　　题 目：The discovery of gravitationally lensed supernova AT 2016jka

　　摘 要：When the light from a distant object passes very near to a foreground galaxy or cluster, gravitational lensing can cause it to appear as multiple images on the sky. If the source is variable, it can be used to constrain the cosmic expansion rate and dark energy models. Lensed supernovae are attractive for this purpose because they have relatively simple photometric behaviour, with well-understood light curve shapes and colours—in contrast to the stochastic variation of quasars. Here, we report the discovery of a multiply imaged supernova, AT 2016jka. It appear in an evolved galaxy at redshift 1.95, gravitationally lensed by a foreground galaxy cluster MACS J0138.0-2155 at redshift 0.338. It is probably a type Ia supernova.

　　报告人：董江（紫台研究生）

　　时 间：2023年12月05日星期二12:15

　　地 点：紫台3-302

欢迎大家参加！