An Introduction to MLE’s

Topics cover:

Consequences of the diffeomorphism invariance of Maximum likelihood estimator s. 1. MLE = rel entropy minimization 1. example: coincides w ith the method of moments/renormalization group. 1. example of a bimmodal uniform distribut ion. 1. One might expect that bimodal distributions become dissimilar to Gaus sians as their displacement goes to zero. 1. accomplished thorugh change of variabiles 1. diff invariance says no. 1. finale: gives an indication that rel entropy is not what we mean in t he presenc e of geometry. motivaitons for interpretability of wasserstien GANs 1. math 1. physics