{"id":"arxiv/2604.03419","url":"https://arxiv.org/abs/2604.03419","tags":["cs.LG","math.CO"],"score":0,"title":"Adaptive Threshold-Driven Continuous Greedy Method for Scalable Submodular Optimization","domain":"cs.LG","source":"arxiv","authors":["Mohammadreza Rostami","Solmaz S. Kia"],"reviews":0,"subject":"cs.LG","summary":"Submodular maximization under matroid constraints is a fundamental problem in combinatorial optimization with applications in sensing, data summarization, active learning, and resource allocation. While the Sequential Greedy (SG) algorithm achieves only a $\\frac{1}{2}$-approximation due to irrevocable selections, Continuous Greedy (CG) attains the optimal $\\bigl(1-\\frac{1}{e}\\bigr)$-approximation via the multilinear relaxation, at the cost of a progressively dense decision vector that forces agents to exchange feature embeddings for nearly every ground-set element. We propose \\textit{ATCG}…","category":"research","comments":0,"disputed":false,"division":"cs","mentions":0,"discussed":0,"published":"2026-04-03","reactions":{}}