Likelihood-free inference in Physical Sciences
Gilles Louppe · Open Repository and Bibliography (University of Liège) · 2018
Refs: Benoit Bovy (xarray-simlab) Cᴀʀʟ Key insightsThe likelihood ratio is suf cient for maximum likelihood estimation. Evaluating the likelihood ratio does not require evaluating the individual likelihoods.Supervised learning indirectly estimates likelihood ratios.