Comparison of classical and Bayesian imaging in radio interferometry

Abstract

CLEAN, the commonly employed imaging algorithm in radio interferometry, suffers from a number of shortcomings: In its basic version, it does not have the concept of diffuse flux, and the common practice of convolving the CLEAN components with the CLEAN beam erases the potential for super-resolution; it does not output uncertainty information; it produces images with unphysical negative flux regions; and its results are highly dependent on the so-called weighting scheme as well as on any human choice of CLEAN masks for guiding the imaging. Here, we present the Bayesian imaging algorithm resolve , which solves the above problems and naturally leads to super-resolution. We take a VLA observation of Cygnus A at four different frequencies and image it with single-scale CLEAN, multi-scale CLEAN, and resolve. Alongside the sky brightness distribution, resolve estimates a baseline-dependent correction function for the noise budget, the Bayesian equivalent of a weighting scheme. We report noise correction factors between 0.4 and 429. The enhancements achieved by resolve come at the cost of higher computational effort.

Geometric Variational Inference

“…Specifically if we consider the geoVI distribution of r given in terms of a generative process and expand it around to first order, we get that Therefore, to first order in y , we get that This correspondence shows that geoVI is a generalization of MGVI in non-linear cases. This is a welcome result, as numerous practical applications [ 29 , 30 , 31 ] have shown that already MGVI provides a sensible approximation to the posterior distribution. On the other hand, it provides further insight in which cases the MGVI approximation remains valid, and when it reaches its limitations.…”

“…This correspondence shows that geoVI is a generalization of MGVI in non-linear cases. This is a welcome result, as numerous practical applications [29][30][31] have shown that already MGVI provides a sensible approximation to the posterior distribution. On the other hand, it provides further insight in which cases the MGVI approximation remains valid, and when it reaches its limitations.…”

Deep images of the Galactic center with GRAVITY

“…In this paper, we present GRAVITY-RESOLVE (G R ), a new imaging code, specifically tailored to GRAVITY observations of the Galactic Center (GC). The tool is based on a Bayesian interpretation of the imaging process and builds upon RESOLVE (Arras et al 2021a(Arras et al , 2018, an imaging tool for radio interferometry developed in the framework of information field theory (Enßlin 2019). In this context, we implemented an instrument model which accounts for all relevant effects in GRAVITY and developed a prior that is specifically designed for the GC and can, for instance, accommodate the variability of Sgr A*.…”

“…In this paper, we present our new imaging code GRAVITY-RESOLVE (G R ) 1 . It builds upon RESOLVE (Arras et al 2021a(Arras et al , 2018, a Bayesian algorithm for radio interferometry, but it is tailored to GC observations with GRAVITY in its measurement equation and its prior model. For exploring the posterior distribution, we employed Metric Gaussian Variational Inference (MGVI, Knollmüller & Enßlin 2019), an algorithm that aims to provide a trade-off between robustness to complicated posterior shapes and applicability to high-dimensional problems.…”

Efficient wide-field radio interferometry response

“…Our implementation has been integrated into the well-known imaging tool wsclean 1 (Offringa et al 2014) since version 2.9, where it can be selected through the -use-wgridder flag, and the imaging A&A 646, A58 (2021) toolkit codex-africanus 2 . Furthermore, the implementation presented here has been used in Arras et al (2020Arras et al ( , 2021, for instance.…”