Interpreting Stellar Spectra with Unsupervised Domain Adaptation

TL;DR

This paper presents an unsupervised domain adaptation method to interpret stellar spectra by transferring knowledge from simulations to observations, improving spectral analysis and feature discovery.

Contribution

It introduces a novel domain transfer pipeline using adversarial autoencoders, disentangled latent spaces, and cycle-consistency for stellar spectroscopic data.

Findings

Improved reconstructed spectra quality.

Ability to discover new spectral features.

Effective transfer from simulated to observed data.

Abstract

We discuss how to achieve mapping from large sets of imperfect simulations and observational data with unsupervised domain adaptation. Under the hypothesis that simulated and observed data distributions share a common underlying representation, we show how it is possible to transfer between simulated and observed domains. Driven by an application to interpret stellar spectroscopic sky surveys, we construct the domain transfer pipeline from two adversarial autoencoders on each domains with a disentangling latent space, and a cycle-consistency constraint. We then construct a differentiable pipeline from physical stellar parameters to realistic observed spectra, aided by a supplementary generative surrogate physics emulator network. We further exemplify the potential of the method on the reconstructed spectra quality and to discover new spectral features associated to elemental abundances.