Unified Projection-Free Algorithms for Adversarial DR-Submodular Optimization

arXiv:2403.10063v1 Announce Type: new
Abstract: This paper introduces unified projection-free Frank-Wolfe type algorithms for adversarial continuous DR-submodular optimization, spanning scenarios such as full information and (semi-)bandit feedback, monotone and non-monotone functions, different constraints, and types of stochastic queries. For every problem considered in the non-monotone setting, the proposed algorithms are either the first with proven sub-linear $\alpha$-regret bounds or have better $\alpha$-regret bounds than the state of the art, where $\alpha$ is a corresponding approximation bound in the offline setting. …

abstract adversarial algorithms arxiv constraints continuous cs.ai cs.cc cs.lg every feedback frank free functions information math.oc optimization paper projection queries stochastic type types