Consistent and accurate estimation of stellar parameters from HARPS-N Spectroscopy using Deep Learning

Consistent and accurate estimation of stellar parameters is great importance for information retrieval in astrophysical research. The span a wide range from effective temperature to rotational velocity. We propose estimate the directly spectral signals coming HARPS-N spectrograph pipeline before any spectrum-processing steps are applied extract 1D spectrum. an attention-based model parameters, which both mean uncertainty through Gaussian distribution. estimated distributions create basis generate data-driven confidence intervals parameters. show that residual networks models can with high accuracy low Signal-to-noise ratio (SNR) compared previous methods. With observation Sun spectrograph, we real observational data.