Decoding RR Lyrae light curves with deep learning for accurate absolute magnitude estimation

Abstract

RR Lyrae stars are essential standard candles for distance measurements in the Milky Way and nearby galaxies. Traditional estimates rely on the Period--Absolute Magnitude--Metallicity relation but are limited by uncertainties in metallicity determinations. We present a deep learning approach that directly predicts absolute magnitudes from RRab and RRc light curves, eliminating the need for metallicity estimates. Our model achieves validation precisions of 0.053 mag and 0.036 mag (approximately 2…