Marginal Bayesian Statistics Using Masked Autoregressive Flows and Kernel Density Estimators with Examples in Cosmology by Harry Bevins et al. on Monday 28 NovemberCosmological experiments often employ Bayesian workflows to deriveconstraints on cosmological and astrophysical parameters from their data. Ithas been shown that these constraints can be combined across different probessuch as Planck and the Dark Energy Survey and that this can be a valuableexercise to improve our understanding of the universe and quantify tensionbetween multiple experiments. However, these experiments are typically plaguedby differing systematics, instrumental effects and contaminating signals, whichwe collectively refer to as `nuisance' components, that have to be modelledalongside target signals of interest. This leads to high dimensional parameterspaces, especially when combining data sets, with > 20 dimensions of which onlyaround 5 correspond to key physical quantities. We present a means by which tocombine constraints from different data sets in a computationally efficientmanner by generating rapid, reusable and reliable marginal probability densityestimators, giving us access to nuisance-free likelihoods. This is possiblethrough the unique combination of nested sampling, which gives us access toBayesian evidences, and the marginal Bayesian statistics code MARGARINE. Ourmethod is lossless in the signal parameters, resulting in the same posteriordistributions as would be found from a full nested sampling run over allnuisance parameters, and typically quicker than evaluating full likelihoods. Wedemonstrate our approach by applying it to the combination of posteriors fromthe Dark Energy Survey and Planck.arXiv: http://arxiv.org/abs/http://arxiv.org/abs/2207.11457v3