In the Euclidean setting, the BFGS method is a well-known quasi-Newton method that has been viewed for
many years as the best quasi-Newton method for solving unconstrained optimization problems. When considering
a cost function defined on a Riemannian manifold, the Euclidean BFGS method cannot be applied directly and
multiple Riemannian versions of BFGS methods have been proposed. Compared to the Euclidean setting, those Riemannian BFGS methods usually have extra cost on an operation, called vector transport. In this presentation, recently developed generic Riemannian BFGS methods are introduced; their differences are highlighted; and an efficient implementation for vector transport is discussed. Numerical experiments are given to demonstrate the
performance of the Riemannian BFGS method with the implementations.