A Primal-Dual Algorithm for Faster Distributionally Robust Optimization

arXiv:2403.10763v1 Announce Type: cross
Abstract: We consider the penalized distributionally robust optimization (DRO) problem with a closed, convex uncertainty set, a setting that encompasses the $f$-DRO, Wasserstein-DRO, and spectral/$L$-risk formulations used in practice. We present Drago, a stochastic primal-dual algorithm that achieves a state-of-the-art linear convergence rate on strongly convex-strongly concave DRO problems. The method combines both randomized and cyclic components with mini-batching, which effectively handles the unique asymmetric nature of the primal and dual problems in DRO. We support …

abstract algorithm art arxiv convergence cs.lg faster linear math.oc optimization practice primal rate risk robust set state stat.ml stochastic type uncertainty