Massively reducing the number of model simulations required in Bayesian synthetic likelihood with little loss of accuracy by exploiting whitening transformations. Led by my Masters student Jacob Priddle and joint with Scott Sisson . @ACEMathStats arxiv.org/abs/1909.04857

Excited to announce our semi-parametric extension for relaxing the normality assumption in Bayesian synthetic likelihood has been accepted by Statistics and Computing Ziwen An @ACEMathStats. Full-text view-only version: rdcu.be/bS67q ArXiv: arxiv.org/abs/1809.05800