Accelerating dark matter halo lensing with generative neural networks

Charles Wilson ( Université de Montréal )

Strong gravitational lensing has proven to be an invaluable probe of the abundance of low mass dark matter halos, a crucial component for resolving the current “Standard Model” of cosmology’s missing satellites problem. Novel approaches for the detection of the extremely faint lensing signal of such halo populations require a large number of simulations of line-of-sight halos, which become unfeasible to produce at scale. I will discuss ways in which machine learning can accelerate such simulations to make them practical. In particular, I will present the results of training generative neural networks to efficiently synthesize deflection angles of line-of-sight halos.