Submodular Maximization via Taylor Series Approximation
Machine Learning
2024-12-17 v1
Abstract
We study submodular maximization problems with matroid constraints, in particular, problems where the objective can be expressed via compositions of analytic and multilinear functions. We show that for functions of this form, the so-called continuous greedy algorithm attains a ratio arbitrarily close to using a deterministic estimation via Taylor series approximation. This drastically reduces execution time over prior art that uses sampling.
Cite
@article{arxiv.2101.07423,
title = {Submodular Maximization via Taylor Series Approximation},
author = {Gözde Özcan and Armin Moharrer and Stratis Ioannidis},
journal= {arXiv preprint arXiv:2101.07423},
year = {2024}
}
Comments
15 pages, 2 figures, to be published in the SIAM International Conference on Data Mining proceedings (SDM 2021)