An Alternative Approach to Sampling: Retrieving Exoplanetary Spectra with Variational Inference and Normalising Flow

Current endeavours in exoplanet characterisation rely on atmospheric retrieval to quantify crucial physical properties of remote exoplanets from observations. However, the scalability and efficiency of the technique are under strain with increasing spectroscopic resolution and forward model complexity. The situation becomes more acute with the recent launch of the James Webb Space Telescope and other upcoming missions. Recent advances in Machine Learning provide optimisation-based Variational Inference as an alternative approach to perform approximate Bayesian Posterior Inference. In this investigation we combined Normalising Flow-based neural network with our newly developed differentiable forward model, Diff-tau, to perform Bayesian Inference in the context of atmospheric retrieval. We demonstrated, with examples from real and simulated spectroscopic data, several advantages of our proposed framework: 1.) Our neural network does not require a large library of spectra, all it takes is a single observation 2.) It is able to produce distributions similar to sampling-based retrieval and 3.) It requires much less forward model computation to converge. Our proposed framework contribute towards the latest development of a neural-powered atmospheric retrieval.  Its flexibility and speed hold the potential to complement sampling-based approaches in large and complex data sets in the future.