Konstantin Karchev

TL;DR: The European Consortium for Astroparticle Theory (EuCAPT) white paper as mentioned in this paper explores upcoming theoretical opportunities and challenges for our field of research with particular emphasis on the possible synergies among different subfields, and the prospects for solving the most fundamental open questions with multi-messenger observations.

TL;DR: A new analysis pipeline is presented that tackles diverse challenges in optical images of galaxy-galaxy strong gravitational lensing systems by bringing together many recent machine learning developments in one coherent approach, including variational inference, Gaussian processes, differentiable probabilistic programming, and neural likelihood-to-evidence ratio estimation.

TL;DR: In this paper , a neural simulation-based inference technique called truncated marginal neural ratio estimation (TMNRE) was proposed to constrain the warm dark matter halo mass function.

TL;DR: In this paper, the authors leverage a range of recent developments in machine learning to develop a new Bayesian strong-lensing image analysis pipeline, which includes a fast, GPU-enabled, end-to-end differentiable image simulator, a statistically principled source model based on a computationally highly efficient approximation to Gaussian processes that also takes into account pixellation, and a scalable variational inference framework that enables simultaneously deriving posteriors for tens of thousands of lens and source parameters and optimising hyperparameters via stochastic gradient descent.

TL;DR: T truncated marginal neural ratio estimation (TMNRE), a form of marginal likelihood-free inference, is applied to Bahamas, a Bayesian hierarchical model for salt parameters and it is verified that TMNRE produces unbiased and precise posteriors for cosmological parameters from up to 100 000 SN Ia.