SM-Net: Learning a Continuous Spectral Manifold from Multiple Stellar Libraries

Omar Anwar, Aaron S. G. Robotham, Luca Cortese, Kevin Vinsen · 2026

Read SM-Net: Learning a Continuous Spectral Manifold from Multiple Stellar Libraries by Omar Anwar, Aaron S. G. Robotham, Luca Cortese, Kevin Vinsen (2026) 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-2603-23899.txt.

Description

Anwar and co-authors present SM-Net, a model that generates stellar spectra from three basic parameters. It takes effective temperature, surface gravity, and metallicity as inputs. Training merges four stellar libraries into one wider, smoother grid. The grid spans temperatures from 2,000 to 190,000 K and a broad wavelength range. SM-Net interpolates smoothly across library boundaries and fills masked flux regions from learned patterns. Test errors stay low across 11,530 spectra, and it makes over 14,000 spectra per second on one GPU.

Catalog record

CREATOROmar Anwar, Aaron S. G. Robotham, Luca Cortese, Kevin Vinsen
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE66803
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.23899

Browse the full library · About the Tower of Babel