All Times ET
Keywords: gravitational lensing, data analysis, high angular resolution, image processing
Modern VLBI (Very Long Baseline Interferometry) arrays now provide us with milli-arcsecond angular resolution images of strong gravitational lens systems at high signal-to-noise ratios. This is a powerful tool for studying astrophysical phenomena, including the evolution of high-redshift galaxies and the particle nature of dark matter. However, properly modelling these data sets poses a computational challenge due to both the large number of data points (as high as 10^10) and the high image resolution needed to pixellate the sky.
In this presentation, I discuss the development of a Bayesian forward-modelling technique for simultaneously reconstructing the source brightness and the lens mass in a self-consistent way from VLBI observations. Our method requires no pre-averaging or other data reduction steps prior to modelling, which sets it apart from previous approaches to this problem. I present the numerical techniques used to overcome the aforementioned computational challenges, as well as the results of tests on simulated data which verify the performance and accuracy of this method.
I also present results obtained using this method from both global VLBI and ALMA (Atacama Large Millimeter Array) observations of strong gravitational lens systems, and finally discuss future directions for both code development and science.