Stronger Approximation Guarantees for Non-Monotone γ-Weakly DR-Submodular Maximization

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Jadav, Hareshkumar, Singh, Ranveer, Aggarwal, Vaneet
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914230329933824
author Jadav, Hareshkumar
Singh, Ranveer
Aggarwal, Vaneet
author_facet Jadav, Hareshkumar
Singh, Ranveer
Aggarwal, Vaneet
contents Maximizing submodular objectives under constraints is a fundamental problem in machine learning and optimization. We study the maximization of a nonnegative, non-monotone $γ$-weakly DR-submodular function over a down-closed convex body. Our main result is an approximation algorithm whose guarantee depends smoothly on $γ$; in particular, when $γ=1$ (the DR-submodular case) our bound recovers the $0.401$ approximation factor, while for $γ<1$ the guarantee degrades gracefully and, it improves upon previously reported bounds for $γ$-weakly DR-submodular maximization under the same constraints. Our approach combines a Frank-Wolfe-guided continuous-greedy framework with a $γ$-aware double-greedy step, yielding a simple yet effective procedure for handling non-monotonicity. This results in state-of-the-art guarantees for non-monotone $γ$-weakly DR-submodular maximization over down-closed convex bodies.
format Preprint
id arxiv_https___arxiv_org_abs_2601_00611
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Stronger Approximation Guarantees for Non-Monotone γ-Weakly DR-Submodular Maximization
Jadav, Hareshkumar
Singh, Ranveer
Aggarwal, Vaneet
Machine Learning
Artificial Intelligence
Computational Complexity
Optimization and Control
Maximizing submodular objectives under constraints is a fundamental problem in machine learning and optimization. We study the maximization of a nonnegative, non-monotone $γ$-weakly DR-submodular function over a down-closed convex body. Our main result is an approximation algorithm whose guarantee depends smoothly on $γ$; in particular, when $γ=1$ (the DR-submodular case) our bound recovers the $0.401$ approximation factor, while for $γ<1$ the guarantee degrades gracefully and, it improves upon previously reported bounds for $γ$-weakly DR-submodular maximization under the same constraints. Our approach combines a Frank-Wolfe-guided continuous-greedy framework with a $γ$-aware double-greedy step, yielding a simple yet effective procedure for handling non-monotonicity. This results in state-of-the-art guarantees for non-monotone $γ$-weakly DR-submodular maximization over down-closed convex bodies.
title Stronger Approximation Guarantees for Non-Monotone γ-Weakly DR-Submodular Maximization
topic Machine Learning
Artificial Intelligence
Computational Complexity
Optimization and Control
url https://arxiv.org/abs/2601.00611