Differentiable Stellar Disk Integration

Inversion of stellar abundance maps based on time series of spectra is a demanding task. The Doppler Imaging (DI) method is successful, but since it is traditionally based on regularized chi-square minimization, it is computationally expensive. The first step to making the progress is to improve the method of generating spectrum for a stellar surface abundances and rotational velocities map. We present a python package that uses Jax (Autograd and XLA package) and is therefore state-of-the-art in terms of computational efficiency.