Stochastic Subspace Descent

David Kozak, Stephen Becker, Alireza Doostan +1

We present two stochastic descent algorithms that apply to unconstrained optimization and are particularly efficient when the objective function is slow to evaluate and gradients a…