Curvature Beyond Positivity: Greedy Guarantees for Arbitrary Submodular Functions

2026-05-08

AI Summary

This work addresses the limitations of classical submodular optimization, which typically assumes non-negativity and monotonicity—assumptions often violated in real-world objectives involving negative values or non-monotone behavior. The paper introduces the first extension of curvature to arbitrary submodular functions and proposes a pruning-based greedy algorithm that leverages multilinear extensions and relaxation techniques to achieve multiplicative approximation guarantees under general combinatorial constraints, even for non-monotone functions with negative values. Theoretical analysis demonstrates that the method unifies the handling of non-monotonicity and negativity through a single curvature parameter, surpassing classical theoretical barriers and yielding an improved approximation ratio for non-monotone settings compared to existing state-of-the-art results. Empirical evaluations on design, coverage, feature selection, and Multi-News paragraph selection tasks confirm the approach’s effectiveness.

Abstract

Submodular functions -- functions exhibiting diminishing returns -- are central to machine learning. When the objective is monotone and non-negative, the greedy algorithm achieves a tight $63\%$ approximation. But many practical objectives incorporate costs that make them negative on some inputs, and all existing multiplicative guarantees require non-negativity. Prior work handles negativity through additive bounds for the special class of decomposable functions and non-monotonicity through partial-monotonicity parameters, but these address each difficulty in isolation and neither extends the classical structural theory. We extend \emph{curvature} -- a parameter measuring how far a function deviates from linearity -- to all submodular functions, handling both non-monotonicity and negativity through a single classical concept. A greedy algorithm with pruning achieves a curvature-controlled multiplicative ratio for \emph{any} submodular function, including those taking negative values -- the first such guarantee beyond monotonicity and non-negativity. In the non-monotone regime $1 \le c_g < 2.2$, the bound strictly beats the best known uniform ratio of $0.401$ (for non-negative $f$), and it recovers the classical $(1-e^{-c_g})/c_g$ guarantee for monotone functions. A multilinear-extension variant extends the framework to general combinatorial constraints via multilinear relaxation. Experiments on cost-penalized experimental design, coverage, feature selection, and a curvature sweep on Multi-News passage selection support the theory.

Problem

• submodular functions
• non-negativity
• monotonicity
• approximation guarantees
• curvature

Innovation

• curvature
• submodular functions
• greedy algorithm
• non-monotone
• negative values