gallifrey : JAX-based Gaussian process structure learning for astronomical time series

C. Boettner , · 2025

Read gallifrey : JAX-based Gaussian process structure learning for astronomical time series by C. Boettner , (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-20394.txt.

Description

Christopher Boettner introduces a software package that learns the shape of noise in star light curves. Gaussian processes are a go to tool for such series, above all in planet hunting. Their weak spot is the kernel, the curve shape a user must pick by hand. Pick wrong, and the fit suffers.

The package, gallifrey, searches kernel structures on its own. It scores shapes by Bayesian rules and samples them with a fresh Monte Carlo scheme. Trials cover star flicker, planet transits, and the spectra of passing atmospheres. Learned kernels matched or beat hand picked ones, with more honest error bars. The author maps next steps toward bigger survey data.

Catalog record

CREATORC. Boettner ,
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE62 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2505-20394

Browse the full library · About the Tower of Babel