Optimization on a Finer Scale: Bounded Local Subgradient Variation Perspective