ResearchTrend.AI
  • Papers
  • Communities
  • Events
  • Blog
  • Pricing
Papers
Communities
Social Events
Terms and Conditions
Pricing
Parameter LabParameter LabTwitterGitHubLinkedInBlueskyYoutube

© 2025 ResearchTrend.AI, All rights reserved.

  1. Home
  2. Papers
  3. 1904.09354
66
104

Submodular Maximization Beyond Non-negativity: Guarantees, Fast Algorithms, and Applications

19 April 2019
Christopher Harshaw
Moran Feldman
Justin Ward
Amin Karbasi
ArXiv (abs)PDFHTML
Abstract

It is generally believed that submodular functions -- and the more general class of γ\gammaγ-weakly submodular functions -- may only be optimized under the non-negativity assumption f(S)≥0f(S) \geq 0f(S)≥0. In this paper, we show that once the function is expressed as the difference f=g−cf = g - cf=g−c, where ggg is monotone, non-negative, and γ\gammaγ-weakly submodular and ccc is non-negative modular, then strong approximation guarantees may be obtained. We present an algorithm for maximizing g−cg - cg−c under a kkk-cardinality constraint which produces a random feasible set SSS such that E[g(S)−c(S)]≥(1−e−γ−ϵ)g(OPT)−c(OPT)\mathbb{E} \left[ g(S) - c(S) \right] \geq (1 - e^{-\gamma} - \epsilon) g(OPT) - c(OPT)E[g(S)−c(S)]≥(1−e−γ−ϵ)g(OPT)−c(OPT), whose running time is O(nϵlog⁡21ϵ)O (\frac{n}{\epsilon} \log^2 \frac{1}{\epsilon})O(ϵn​log2ϵ1​), i.e., independent of kkk. We extend these results to the unconstrained setting by describing an algorithm with the same approximation guarantees and faster O(nϵlog⁡1ϵ)O(\frac{n}{\epsilon} \log\frac{1}{\epsilon})O(ϵn​logϵ1​) runtime. The main techniques underlying our algorithms are two-fold: the use of a surrogate objective which varies the relative importance between ggg and ccc throughout the algorithm, and a geometric sweep over possible γ\gammaγ values. Our algorithmic guarantees are complemented by a hardness result showing that no polynomial-time algorithm which accesses ggg through a value oracle can do better. We empirically demonstrate the success of our algorithms by applying them to experimental design on the Boston Housing dataset and directed vertex cover on the Email EU dataset.

View on arXiv
Comments on this paper