A variational approach to stochastic minimization of convex functionals

Miroslav Bacak

Stochastic methods for minimizing a convex integral functional, as initiated by Robbins and Monro in the early 1950s, rely on the evaluation of a gradient (or subgradient if the fu…