Short description: We use amortized neural posterior estimation of the temperature distribution in fast rotators. The posterior distribution is approximated with conditional normalizing flows conditioned on spectroscopic observations. The conditioning is achieved through the use of Transformer encoders, which can deal with arbitrary wavelength sampling and rotation phases.
Title: Approximate Bayesian neural Doppler imaging
Short description: We use amortized neural posterior estimation of the temperature distribution in fast rotators. The posterior distribution is approximated with conditional normalizing flows conditioned on spectroscopic observations. The conditioning is achieved through the use of Transformer encoders, which can deal with arbitrary wavelength sampling and rotation phases.
Link to paper: https://ui.adsabs.harvard.edu/abs/2022A%26A...658A.162A/abstract