New fractional regularization framework enhances sparse signal recovery

Researchers have introduced a novel unified fractional regularization framework designed for sparse signal recovery using the $\ell_1/\ell_p^q$ model. This framework establishes an equivalence between first-order stationary points of the $\ell_1/\ell_p^q$ formulation and the subtractive $\ell_1 - \alpha \ell_p$ model. The paper also presents a new recovery condition under the Restricted Isometry Property (RIP) and details a majorization-minimization (MM) algorithm for solving the problem, demonstrating superior performance in numerical experiments for MRI reconstruction. AI

影响 Introduces a new mathematical framework and algorithm for sparse signal recovery, potentially improving performance in applications like MRI reconstruction.

排序理由 This is a research paper detailing a new theoretical framework and algorithm for sparse signal recovery.

在 arXiv cs.LG 阅读 →

• arXiv
• $\ell_1 - \alpha \ell_p$
• $\\ell_1/\\ell_p^q$
• Kurdyka-Lojasiewicz
• Restricted Isometry Property
• RIP

• 论文
• 其他

AI 生成摘要 · Google Gemini · 来自 1 个来源。 我们如何撰写摘要 →