These are algorithms for convex optimization whereas stochastic gradient descent and saddle free newton are generally used for nonconvex optimization (neural nets in particular).