English

Minimum Robust Multi-Submodular Cover for Fairness

Data Structures and Algorithms 2020-12-16 v1

Abstract

In this paper, we study a novel problem, Minimum Robust Multi-Submodular Cover for Fairness (MinRF), as follows: given a ground set VV; mm monotone submodular functions f1,...,fmf_1,...,f_m; mm thresholds T1,...,TmT_1,...,T_m and a non-negative integer rr, MinRF asks for the smallest set SS such that for all i[m]i \in [m], minXrfi(SX)Ti\min_{|X| \leq r} f_i(S \setminus X) \geq T_i. We prove that MinRF is inapproximable within (1ϵ)lnm(1-\epsilon)\ln m; and no algorithm, taking fewer than exponential number of queries in term of rr, is able to output a feasible set to MinRF with high certainty. Three bicriteria approximation algorithms with performance guarantees are proposed: one for r=0r=0, one for r=1r=1, and one for general rr. We further investigate our algorithms' performance in two applications of MinRF, Information Propagation for Multiple Groups and Movie Recommendation for Multiple Users. Our algorithms have shown to outperform baseline heuristics in both solution quality and the number of queries in most cases.

Keywords

Cite

@article{arxiv.2012.07936,
  title  = {Minimum Robust Multi-Submodular Cover for Fairness},
  author = {Lan N. Nguyen and My T. Thai},
  journal= {arXiv preprint arXiv:2012.07936},
  year   = {2020}
}