Let be two -by- Hermitian matrices with eigenvalues and , respectively. \iffalse There are two kinds of perturbation bounds on : , where is the largest singular value of , regardless of ’s spectral distributions, and , where is the minimum gap between ’s spectra. \end{enumerate} Bounds of the first kind overestimate the changes when while those of the second kind may blow up when is too tiny. \fi Denote by the spectral norm of the matrix , and the spectral gap between the spectra of and . It is shown that which improves all the existing results. Similar bounds are obtained for singular values of matrices under block perturbations.

生成讲解博客的中文版本

生成AI智能校验的中文版本