Targeted likelihood-free inference for dark matter searches in strong lensing images.

Presentation given by Christoph Weniger, associate professor at GRAPPA, the centre of excellence for GRavitation AstroParticle Physics Amsterdam, and the Institute of Physics at the University of Amsterdam.
Abstract: Images of gravitationally lensed galaxies provide information about the small scale structure of dark matter. I will present a new analysis pipeline for strong lensing images that combines a range of ML techniques, including probabilistic differentiable programming, variational inference and neural ratio estimation."
Link to paper: arxiv.org/abs/2010.07032