Using LLMs to Classify Astrophysical Transients

Can LLMs be used to classify astrophysical transients?

A recent paper in nature astronomy showed that using only 15 images, Gemini can achieve 93% accuracy in transient classification.

We discuss in detail the research paper and highlights some interesting findings.

-----------------------------------------------------------------

Main paper discussed in the video:

"Textual interpretation of transient image classifications from large language models": F. Stoppa. T. Bulmus et al. arXiv: 2510.06931

Other paper shown:

"Transient-optimized real-bogus classification with Bayesian convolutional
neural networks – sifting the GOTO candidate stream" : T.L. Killestein et al., MNRAS 503, 4838–4854 (2021)

----------------------------------------------

00:30 - Transients
01:30 - Difference Imaging
04:25 - Real or Bogus Transients?
07:15 - Convolution is Great, but..
12:06 - Textual Interpretation of Images; via LLMs !!!

Видео Using LLMs to Classify Astrophysical Transients канала BlackHoles-to-BackProp