KL property of exponent $1/2$ for zero-norm composite quadratic functions
Yuqia Wu, Shaohua Pan, Shujun Bi · arXiv (Cornell University) · 2018
This paper is concerned with a class of zero-norm regularized and constrained composite quadratic optimization problems, which has important applications in the fields such as sparse eigenvalue problems, sparse portfolio problems, and nonnegative matrix factorizations. For this class of nonconvex and nonsmooth problems, we establish the KL property of exponent 1/2 of its objective function under a suitable assumption, and provide some examples to illustrate that the assumption holds.