English

Sparsity via Sparse Group $k$-max Regularization

Machine Learning 2024-02-14 v1 Machine Learning

Abstract

For the linear inverse problem with sparsity constraints, the l0l_0 regularized problem is NP-hard, and existing approaches either utilize greedy algorithms to find almost-optimal solutions or to approximate the l0l_0 regularization with its convex counterparts. In this paper, we propose a novel and concise regularization, namely the sparse group kk-max regularization, which can not only simultaneously enhance the group-wise and in-group sparsity, but also casts no additional restraints on the magnitude of variables in each group, which is especially important for variables at different scales, so that it approximate the l0l_0 norm more closely. We also establish an iterative soft thresholding algorithm with local optimality conditions and complexity analysis provided. Through numerical experiments on both synthetic and real-world datasets, we verify the effectiveness and flexibility of the proposed method.

Keywords

Cite

@article{arxiv.2402.08493,
  title  = {Sparsity via Sparse Group $k$-max Regularization},
  author = {Qinghua Tao and Xiangming Xi and Jun Xu and Johan A. K. Suykens},
  journal= {arXiv preprint arXiv:2402.08493},
  year   = {2024}
}

Comments

7 pages, accepted to American Control Conference 2024

R2 v1 2026-06-28T14:47:23.229Z