On the Schatten \(p\)-quasi-norm minimization for low-rank matrix recovery (Q2659733)

Let \(\Omega\) be a given subset of the index set \(\{(i,j),i= 1,\dots,m,j=1,\dots,n\}\). Suppose that the entries of a matrix \(M\) of the size \(m\times n\) with indices in \(\Omega\) are given. The so-called matrix completion problem seeks a matrix \(X\) of the size \(m\times n\) with smallest rank such that the entries of \(X\) in \(\Omega\) are the same as the entries of \(M\) in \(\Omega\). This is closely related to the low-rank matrix recovery problem asking for \(\min_{X\in\mathbb{R}^{m\times n}}\text{rank}(X)\) such that \(\mathcal{A}(X)=\mathcal{A}(M)\), where \(\mathcal{A}: \mathbb{R}^{m\times n} \to \mathbb{R}^\ell\) is a linear mapping. The authors prove some results about inequalities in the Schatten \(p\)-quasi-norm of matrices and then use them for establishing a sufficient condition when the Schatten \(p\)-quasi-norm minimization can be used for low rank matrix recovery.