Gelfand numbers of embeddings of Schatten classes (Q2049959)

For \(0< p \leq \infty\), the Schatten \(p\)-class \(\mathcal{S}_p\) is the Banach space of all compact operators on a Hilbert space \(H\) which have singular values in the sequence space \(\ell_p\). The most important special cases are given by the trace class operators \((p =1)\), Hilbert-Schmidt operators \((p =2)\), and compact operators \((p =\infty)\). Given \(N \in \mathbb N\), the Schatten \(p\)-class of all operators (\( = N \times N\)-matrices) on the \(N\)-dimensional Hilbert space \(\ell_2^N\) is denoted by \(\mathcal{S}^N_p\). In geometric functional analysis, the study of Schatten classes has a long tradition, and the structure theory of these operators has been investigated intensively over the last seven decades. Today, Schatten classes provide the mathematical framework to modern applied mathematics around low-rank matrix recovery (the non-commutative analog of the classical compressed sensing approach), and are of fundamental interest in quantum information theory. Since operators in Schatten classes are compact by definition, it seems natural to quantify the degree of compactness of such operators through their corresponding sequence of so-called Gelfand numbers, which form one of the most crucial concepts in modern approximation and complexity theory. Given a bounded operator \(T: X \to Y\) between two (quasi-)Banach spaces \(X\) and \(Y\), the \(n\)th Gelfand number is defined by \[c_n(T) = \inf \big\{ \|T|_F \| : F \subset X, \ \text{codim}\, F < n \big\}\,.\] The literature offers several articles analysing the decay of the sequence of Gelfand numbers \(c_n(\mathcal{S}^N_p \hookrightarrow \mathcal{S}^N_q)\), where \(0< p,q\leq \infty\) and \(1 \leq n \leq N^2\). Among others, partial results were given in [\textit{B. Carl} and \textit{A. Defant}, J. Approx. Theory 88, No. 2, 228--256 (1997; Zbl 0892.46075); \textit{J. A. Chávez-Domínguez} and \textit{D. Kutzarova}, J. Math. Anal. Appl. 427, No. 1, 320--335 (2015; Zbl 1331.41036); \textit{A. Hinrichs} and \textit{C. Michels}, Suppl. Rend. Circ. Mat. Palermo (2) 76, 395--411 (2005; Zbl 1151.47029)]. This interesting paper complements the known results on Gelfand numbers of natural embeddings between \(N\)-dimensional Schatten classes. It provides asymptotically sharp bounds for almost all missing regimes, and in fact there were many such cases left. To solve such a difficult `puzzle' one needs deep knowledge in geometric functional analysis. But not only this, in order to reach an almost complete picture various elements and techniques of analytic, geometric and probabilistic flavour have to be combined and extended in a nontrivial way.