Gelfand numbers of Schatten class embeddings in the quasi-Banach range
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Mathematics. The day in mathematics: new theorems, new proofs, and what they open.
Ada: We're looking at the paper: "Gelfand numbers of Schatten class embeddings in the quasi-Banach range".
Emmy: Here's the gist of the paper. Gelfand numbers of Schatten class embeddings in the quasi-Banach range proves Conjecture 1.4 of Prochno and Strzelecki by establishing a complete asymptotic description for Gelfand numbers in the range where previous results were incomplete The main result establishes the conjectured lower estimate for any fixed proportion of the ambient dimension, combining known upper estimates to provide an exact asymptotic order Context and Motivation The Schatten classes form a classical family of operator ideals in functional analysis, where the Schatten p-class Sp consists of all compact operators on a Hilbert space whose sequence of singular values belongs to lp.
Ada: The geometry of these spaces is not determined by the corresponding sequence-space results, as the geometry of Schatten unit balls is not determined by the corresponding sequence-space results.
Emmy: The study of Gelfand numbers quantifies how small a section of the domain unit ball can be made in the target norm, which requires understanding singular value structure and comparison of different Schatten quasi-norms.
Ada: This distinction between commutative and noncommutative geometry is central to the problem.
Title and authors: Ada: Let's talk about the title again, "Gelfand numbers of Schatten class embeddings in the quasi-Banach range." It’s precise because it points directly to the core objects they are studying: Gelfand numbers of mappings between Schatten classes.
Emmy: And it specifies the range, which is where things get tricky for these operator ideals, focusing on when p is less than or equal to one and q starts at two and goes up to infinity.
Kolya: From a structural view, this paper seems interested in how the singular value structure of matrices dictates how well they embed into different norms.
Ada: They are using those Gelfand numbers, which measure the size of a section of the domain unit ball in the target norm, to probe that singular value structure across different quasi-norms.
Emmy: It connects commutative and noncommutative geometry here because these embeddings don't settle all the questions about matrices by themselves; you need to look at low-codimensional matrix subspaces.
Kolya: That’s interesting, because it suggests that understanding how these spaces interact is central to the whole problem of operator embeddings.
The paper's summary: Ada: The main finding they present is Theorem A, which asserts the conjectured lower estimate for the Gelfand numbers in that range of p and q.
Emmy: They show there exist constants such that the Gelfand number of mapping from S N p to S N q is bounded below by a term involving p and q, like c * p to infinity times something with an exponent related to one over two minus one over p.
Kolya: That lower estimate is what completes the asymptotic description for all indices in the range zero less than or equal to p and two less than or equal to q infinity <ref:2611.08448#pg1>.
Ada: And they combine that lower estimate with the known upper estimates from other work, specifically those related to S N two which implies that they get an exact asymptotic order for these embeddings <ref:2611.08448#pg1>.
Emmy: So in simple terms, they've managed to prove the conjectured lower bound and matched it up with the known upper bounds.
Kolya: It’s a complete all-index description by combining their result with the estimates from Hinrichs, Prochno and Vybíral, and also with Prochno and Strzelecki.
The paper's improvements: Ada: One of the big improvements they highlight is that they managed to prove this lower estimate using an elementary projection argument combined with a reduction to rectangular matrices.
Emmy: That method involves considering orthogonal projections onto subspaces of rectangular matrices and then using an averaging argument to get a lower bound on the norm comparison.
Kolya: I'm curious about the specific lemma they use, Lemma three point one, which provides that subspace estimate for a subspace E in M N r of dimension d greater than or equal to one <ref:2611.08448#pg1>.
Ada: And Proposition three point two gives them a lower bound for the Gelfand numbers by maximizing over an index r, asserting that the Gelfand number is at least something involving n p to infinity times something with one over p <ref:2611.08448#pg1>.
Emmy: The real structural improvement is that this proof doesn't require convexity of the Schatten unit ball or quasi-Banach duality, which is a significant limitation for other approaches.
Conclusion: Ada: To wrap things up, the main implication here is that they’ve provided a matching lower bound to close a gap in the description of Gelfand numbers for indices below two.
Emmy: This work shows that this lower estimate holds even when p is less than one, which was previously an open question for these embeddings between Schatten classes.
Kolya: I think what matters most is that they established this matching lower bound by carefully selecting the rectangular dimension r to balance the intersection dimension against rank-dependent norm estimates.
Ada: That selection of r resulted in a constant cp eta that is independent of q, which is a nice piece of information for applications.
Emmy: And since they established the upper bound using known results for S N two this work gives us the complete asymptotic description by matching both sides <ref:2611.08448#pg1>.
University of Passau
math.FA
Submitted: 2026-11-05
Updated: 2026-11-06
Comments: 11 page, 1 figure
License: http://creativecommons.org/licenses/by/4.0/
Importance score: 83/100
The gist: Gelfand numbers of Schatten class embeddings in the quasi-Banach range proves Conjecture 1.4 of Prochno and Strzelecki by establishing a complete asymptotic description for Gelfand numbers in the…
Key concepts
- Schatten p-class (Sp)
- This is a class of compact operators on Hilbert spaces. An operator belongs to Sp if its singular values belong to the sequence space lp. This concept defines the basic structure of the Schatten classes studied in this paper, which are central to understanding operator geometry.
- Gelfand numbers
- These numbers quantify how small a section of a domain unit ball can be made when measured in a target norm. They are crucial for comparing different Schatten quasi-norms and require an understanding of the singular value structure of the operators involved.
- Quasi-Banach range
- This refers to the specific range of indices studied, namely 0 < p <= 1 and 2 <= q <= infinity. The paper focuses on proving these estimates in this particular setting, where previous results were incomplete.
Terminology
Summary
Gelfand numbers of Schatten class embeddings in the quasi-Banach range proves Conjecture 1. p. 0 Gelfand numbers of Schatten class embeddings
in the quasi-Banach range
Joscha Prochno*
Abstract
We prove Conjecture 1. 4 of Prochno and Strzelecki [J.Where the paper says it
Context and Motivation
The Schatten classes form a classical family of operator ideals in functional analysis, where the Schatten p-class Sp consists of all compact operators on a Hilbert space whose sequence of singular values belongs to lp. The geometry of these spaces is not determined by the corresponding sequence-space results, as the geometry of Schatten unit balls is not determined by the corresponding sequence-space results. p. 0 Nevertheless, the geometry of Schatten classes
is not determined by the corresponding sequence-space results. Building on the results of Kwapie´n
and Pełczy´nski [21], Gordon and Lewis [8] proved that, on an infinite-dimensional Hilbert space, the
Banach spaces Sp, 1 ≤ p ≤ ∞, fail to have local unconditional structure whenever p ̸= 2. Geometrically, these numbers measure how small a section of the domain unit ball
can be made in the target norm. Their determination therefore requires understanding the singularvalue structure of matrices in subspaces of prescribed codimension and how it governs the comparison of different Schatten quasi-norms. Geometrically, these numbers measure how small a section of the domain unit ball
can be made in the target norm. Their determination therefore requires understanding the singularvalue structure of matrices in subspaces of prescribed codimension and how it governs the comparison of different Schatten quasi-norms. Geometrically, these numbers measure how small a section of the domain unit ball
can be made in the target norm. Their determination therefore requires understanding the singularvalue structure of matrices in subspaces of prescribed codimension and how it governs the comparison of different Schatten quasi-norms. This brings the distinction between commutative and noncommutative geometry directly into the problem: sharp estimates for the embeddings ℓ
N
p
,→ ℓ
N
q do
not, by themselves, settle the corresponding questions for matrices. The geometry of low-codimensional matrix subspaces that underlies the study of Gelfand numbers
also plays a central role in low-rank matrix recovery and completion, where Schatten norms and
quasi-norms play a role analogous to that of sequence norms and quasi-norms in sparse recovery;
see [1, 3, 5, 20, 24].Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Main Theorem and Results
Theorem A establishes the conjectured lower estimate for 0 < p ≤ 1 and 2 ≤ q ≤ ∞, asserting that there exist constants such that cp,η min½1,Nn¾1/p−1/2≤ cn(S Np,→ S Nq)≤Cp min½1,Nn¾1/p−1/2. This lower estimate is the contribution of this paper and completes the asymptotic description for all indices in the range 0 < p ≤ 1 and 2 ≤ q ≤ ∞. p. 1 The purpose of the present paper is to prove the lower estimate proposed in [22, Conjecture 1. Together
with the previously known bounds, including the boundary cases p = 1 and q = 2, this completes the
asymptotic description of the Gelfand numbers for all indices in the range 0 < p ≤ 1 and 2 ≤ q ≤ ∞. However, the corresponding description of Gelfand numbers
proved less complete. In [14], the authors obtained asymptotically sharp estimates in a wide range
of parameters, extending classical sequence-space results and complementing the earlier bounds.Where the paper says it
Where the paper says it
Proof Strategy
The proof employs an elementary projection argument combined with a reduction to rectangular matrices. The core idea involves considering the orthogonal projections onto subspaces of rectangular matrices and using an averaging argument to yield a lower bound on the norm comparison. p. 3 An averaging argument,
combined with the rank bound for N ×r matrices, then yields
sup
A∈F\{0}
∥A∥∞
∥A∥p
≥
s
d
N r
r
1/2−1/p
, 0 < p ≤ 2. In particular, when F occupies a fixed proportion of the rectangular matrix space, this lower bound
has order r
1/2−1/p
. An averaging argument,
combined with the rank bound for N ×r matrices, then yields
sup
A∈F\{0}
∥A∥∞
∥A∥p
≥
s
d
N r
r
1/2−1/p
, 0 < p ≤ 2. In particular, when F occupies a fixed proportion of the rectangular matrix space, this lower bound
has order r
1/2−1/p
. Intersecting it with
a subspace of codimension less than n leaves a subspace of dimension at least N r −n +1. We keep r
free because the resulting lower bound depends on r , and a suitable choice of r later on will balance
the intersection dimension against the rank-dependent norm estimate. Intersecting it with
a subspace of codimension less than n leaves a subspace of dimension at least N r −n +1. We keep r
free because the resulting lower bound depends on r , and a suitable choice of r later on will balance
the intersection dimension against the rank-dependent norm estimate.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Key Lemmas and Bounds
Lemma 3.1 provides the subspace estimate for a subspace E ⊆ MN,r of dimension d ≥ 1, stating that sup A∈E>0∥A>∞∥A>p≥ p d p N r 1/p. p. 5 Restricting to a smaller subspace cannot increase the restriction norm. Hence the infimum
may be taken over subspaces of dimension exactly N
2 −n +1, and
cn(S
N
p
,→ S
N
q
) = inf
F⊆MN
dimF=N
2−n+1
sup
A∈F\{0}
∥A∥q
∥A∥p
. As already mentioned, the purpose of this paper is to determine the Gelfand numbers of S
N
p
,→ S
N
q
in the range 0 < p ≤ 1 and 2 ≤ q ≤ ∞. Hinrichs, Prochno and Vybíral [14, Proposition 4. Theory 277 (2022), 105736] on
Gelfand numbers of embeddings between finite-dimensional Schatten classes. For 0 < p ≤ 1 and
indices bounded away from the ambient dimension by a fixed proportion, we obtain the conjectured lower bound, uniformly in the target exponent 2 ≤ q ≤ ∞. Empty ranges are omitted. Combining the three cases proves the lower estimate in (3); one may take
cp,η = min½
1,
3
−α
p
2
,
p
η2
−α
¾
.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Upper Bounds and Asymptotics
The upper estimate is known and follows from the corresponding bound with target space S N2, as q ≥ 2 implies∥A>q∥A>p ≤∥A>q∥A>2. The upper estimate is proven using the known bound cn(S Np,→ S N2) ≤ Cp min½1,Nn¾1/p−1/2. p. 0 For 0 < p ≤ 1 and
indices bounded away from the ambient dimension by a fixed proportion, we obtain the conjectured lower bound, uniformly in the target exponent 2 ≤ q ≤ ∞. Together with the known upper
estimate obtained by Hinrichs, Prochno and Vybíral [Math. The boundary of the two highlighted regions is not part
of the open cases: sharp asymptotic estimates were already known for p = 1 with 2 < q ≤ ∞, for q = 2
with 0 < p < 1, for p = 2 with 2 < q ≤ ∞, and along the diagonal p = q > 2. 1 (The complete range of indices). The estimates in the second range were already sharp. The contribution of Theorem A is the missing
matching lower bound in the first range for 0 < p < 1 and q > 2.Where the paper says it
Where the paper says it
Where the paper says it
Conclusion
The paper successfully proves the lower estimate in Theorem A by carefully selecting the rectangular dimension r to balance the intersection dimension against rank-dependent norm estimates. The resulting constant cp,η is independent of q, and the upper bound is established using known results for S N2. p. 8 (10)
In particular, this constant is independent of q. For the upper estimate, we use the known bound
cn(S
N
p
,→ S
N
2
) ≤ Cp min½
1,
N
n
¾α
, 1 ≤ n ≤ N
2
; (11)
see [22, Proposition 3.Where the paper says it
2 by providing a matching lower bound. The proof relies on arguments that do not require convexity of the Schatten unit ball or quasiBanach duality, applicable even for 0 < p < 1. p. 4 The
initial range n ≤ N follows from a rank-one argument, while the passage to all q ≥ 2 uses the trivial
estimate ∥A∥q ≥ ∥A∥∞. The argument uses neither convexity of the Schatten unit ball nor quasiBanach duality and therefore applies without modification when 0 < p < 1. The
initial range n ≤ N follows from a rank-one argument, while the passage to all q ≥ 2 uses the trivial
estimate ∥A∥q ≥ ∥A∥∞. The argument uses neither convexity of the Schatten unit ball nor quasiBanach duality and therefore applies without modification when 0 < p < 1.Where the paper says it
Where the paper says it
References
[1] E. J. Candès and B. p. 8 References
[1] E. Candès and B. Candès and B. Exact matrix completion via convex optimization.Where the paper says it
Where the paper says it
[2] B. Carl and A. p. 1 For Gelfand numbers, fundamental contributions are due to Gluskin [7] and Garnaev and
Gluskin [6], whereas the entropy numbers of embeddings between finite-dimensional ℓp-spaces
were studied by Schütt [30] and subsequently in [4, 32, 10, 19]. In the non-commutative setting,
related approximation quantities for embeddings between Schatten classes had already been investigated by Carl and Defant [2] and Hinrichs and Michels [12], in addition to the work of ChávezDomínguez and Kutzarova [3] discussed above. Carl and A. Asymptotic estimates for approximation quantities of tensor product
identities.Where the paper says it
Where the paper says it
[3] J. A. Chávez-Domínguez and D. p. 1 In particular, the range 0 < p < 1 is relevant to non-convex recovery by Schatten quasi-norm minimization. The connection with Gelfand numbers is more than a formal analogy: Chávez-Domínguez and Kutzarova [3] characterized the stability of such recovery procedures
in terms of Gelfand widths of suitable Schatten class embeddings. Chávez-Domínguez and D. Stability of low-rank matrix recovery and its connections to Banach space geometry.Where the paper says it
Where the paper says it
[4] D. E. Edmunds and H. p. 8 , 427(1):320–335, 2015. Edmunds and H.Where the paper says it
[5] S. Foucart and H. p. 9 Foucart and H. A mathematical introduction to compressive sensing. A mathematical introduction to compressive sensing. Applied and
Numerical Harmonic Analysis.Where the paper says it
Where the paper says it
[6] A. Y. Garnaev and E. p. 1 The study of general approximation quantities for finite-dimensional sequence spaces has a long
history. For Gelfand numbers, fundamental contributions are due to Gluskin [7] and Garnaev and
Gluskin [6], whereas the entropy numbers of embeddings between finite-dimensional ℓp-spaces
were studied by Schütt [30] and subsequently in [4, 32, 10, 19]. The widths of a Euclidean ball. Soviet Math.Where the paper says it
Where the paper says it
[7] E. D. Gluskin, Norms of random matrices and diameters of finite-dimensional sets, Mat. p. 9 , 30:200–204,
1984. Norms of random matrices and diameters of finite-dimensional sets.Where the paper says it
[8] Y. Gordon and D. p. 0 Nevertheless, the geometry of Schatten classes
is not determined by the corresponding sequence-space results. Building on the results of Kwapie´n
and Pełczy´nski [21], Gordon and Lewis [8] proved that, on an infinite-dimensional Hilbert space, the
Banach spaces Sp, 1 ≤ p ≤ ∞, fail to have local unconditional structure whenever p ̸= 2. Gordon and D. Absolutely summing operators and local unconditional structures.Where the paper says it
Where the paper says it
[9] A Grothendieck, Sur une notion de produit tensoriel topologique d’espaces vectoriels topologiques, et une classe remarquable d’espaces vectoriels liée à cette notion. C. R. Acad. Sci. Paris, 233:1556–1558, 1951
[10] O Guédon and A E Litvak, Euclidean projections of a p-convex body, In Geometric aspects of functional analysis, volume 1745 of Lecture Notes in Math., pages 95–108. p. 9 In Geometric aspects of
functional analysis, volume 1745 of Lecture Notes in Math. , pages 95–108.Where the paper says it
[11] O Guédon and G Paouris, Concentration of mass on the Schatten classes, Ann. Inst. H Poincaré Probab. p. 9 Concentration of mass on the Schatten classes. Poincaré
Probab. Concentration of mass on the Schatten classes. Poincaré
Probab.Where the paper says it
Where the paper says it
[12] A Hinrichs and C Michels, Approximation numbers of inclusions between Schatten classes, Rend. Circ. Mat. Palermo (2) Suppl. p. 9 Approximation numbers of inclusions between Schatten classes. Palermo (2) Suppl.Where the paper says it
[13] A Hinrichs, J Prochno, and J Vybíral, Entropy numbers of embeddings of Schatten classes, J. Funct. Anal., 273(10):3241–3261, 2017
[14] A Hinrichs, J Prochno, and J Vybíral, Gelfand numbers of embeddings of Schatten classes, Math. Ann., 380(3–4):1563–1593, 2021
[15] Z Kabluchko, J Prochno, and C Thäle, Exact asymptotic volume and volume ratio of Schatten unit balls, J. Approx. Theory, 257:105457, 2020
[16] Z Kabluchko, J Prochno, and C Thäle, Intersection of unit balls in classical matrix ensembles, Israel J. Math., 239(1):129–172, 2020
[17] Z Kabluchko, J Prochno, and C Thäle, Sanov-type large deviations in Schatten classes, Ann. Inst. Henri Poincaré Probab. p. 9 Sanov-type large deviations in Schatten classes. Henri Poincaré Probab.Where the paper says it
[18] H König, M Meyer, and A Pajor, The isotropy constants of the Schatten classes are bounded, Math. Ann., 312(4):773–783, 1998
[19] T Kühn, A lower estimate for entropy numbers, J. Approx. Theory, 110(1):120–124, 2001
[20] C Kümmerle and J Sigl, Harmonic mean iteratively reweighted least squares for low-rank matrix recovery, J. Mach. Learn. Res., 19:Paper no. p. 1 One concerns embeddings from S
N
p
, with 0 < p < 1, into S
N
q
,
with 2 < q ≤ ∞: sharp upper bounds were available, but a matching lower estimate throughout the
range of indices bounded away from N
2 by a fixed proportion had not been established. The purpose of the present paper is to prove the lower estimate proposed in [22, Conjecture 1.Where the paper says it
[21] S Kwapie´n and A Pełczy´nski, The main triangle projection in matrix spaces and its applications, Studia Math. p. 9 The main triangle projection in matrix spaces and its applications. Studia Math.Where the paper says it
[22] J Prochno and M Strzelecki, Approximation, Gelfand, and Kolmogorov numbers of Schatten class embeddings, J. Approx. Theory, 277:105736, 2022
[23] J Radke and B.-H Vritsiou, On the thin-shell conjecture for the Schatten classes, Ann. p. 9 Radke and B. On the thin-shell conjecture for the Schatten classes.Where the paper says it
[24] A Rohde and A B Tsybakov, Estimation of high-dimensional low-rank matrices, Ann. Statist., 39(2):887–930, 2011
[25] A F Ruston, On the Fredholm theory of integral equations for operators belonging to the trace class of a general Banach space, Proc. London Math. p. 0 For 0 < p ≤ 1 and
indices bounded away from the ambient dimension by a fixed proportion, we obtain the conjectured lower bound, uniformly in the target exponent 2 ≤ q ≤ ∞. Together with the known upper
estimate obtained by Hinrichs, Prochno and Vybíral [Math.Where the paper says it
[26] R Schatten, The cross-space of linear transformations, Ann. of Math.(2), 47:73–84, 1946
[27] R Schatten and J von Neumann, The cross-space of linear transformations. II, Ann. of Math.(2), 47:608–630, 1946
[28] R Schatten and J von Neumann, The cross-space of linear transformations. III, Ann. of Math.(2), 49:557–582, 1948
[29] C Schütt, Entropy numbers of diagonal operators between symmetric Banach spaces, J. Approx. Theory., 40(2):121–128, 1984
[30] N Tomczak-Jaegermann, The moduli of smoothness and convexity and the Rademacher averages of trace classes Sp (Sp (Sp (Sp (Sp(Sp(Sp(Sp(NTomczak-Jaegermann, The moduli of smoothness and convexity and the Rademacher averages of trace classes Sp (1≤p<∞). Studia Math., 50:163–182, 1974
[31] H Triebel, Fractals and spectra, volume 91 of Monographs in Mathematics. Birkhäuser Verlag, Basel, 1997
[32] Related to Fourier analysis and function spaces. C R Acad Sci Paris, 233:1556–1558, 1951
--- Page 3 ---
The Schatten classes form one of the classical families of operator ideals in functional analysis, where the Schatten p-class Sp consists of all compact operators on a Hilbert space whose sequence of singular values belongs to lp. The geometry of these spaces is not determined by the corresponding sequence-space results. p. 0 Nevertheless, the geometry of Schatten classes
is not determined by the corresponding sequence-space results. Building on the results of Kwapie´n
and Pełczy´nski [21], Gordon and Lewis [8] proved that, on an infinite-dimensional Hilbert space, the
Banach spaces Sp, 1 ≤ p ≤ ∞, fail to have local unconditional structure whenever p ̸= 2. Nevertheless, the geometry of Schatten classes
is not determined by the corresponding sequence-space results. Building on the results of Kwapie´n
and Pełczy´nski [21], Gordon and Lewis [8] proved that, on an infinite-dimensional Hilbert space, the
Banach spaces Sp, 1 ≤ p ≤ ∞, fail to have local unconditional structure whenever p ̸= 2.Where the paper says it
Where the paper says it
Pruning redundant reasoning tokens
The Schatten classes form one of the classical families of operator ideals in functional analysis, where the Schatten p-class Sp consists of all compact operators on a Hilbert space whose sequence of singular values belongs to lp. The geometry of these spaces is not determined by the corresponding sequence-space results. The study of Gelfand numbers quantifies how small a section of the domain unit ball can be made in the target norm, which requires understanding singular value structure and comparison of different Schatten quasi-norms.
How the training signal is built
The proof combines an elementary projection argument with a reduction to rectangular matrices. The core idea involves considering the orthogonal projections onto subspaces of rectangular matrices and using an averaging argument to yield a lower bound on the norm comparison. Specifically, for a linear subspace F ⊆ R NxN of codimension less than n, the estimate leads to sup A∈F>0∥A>q∥A>p≥ s d N r r 1/2−1/p, where r is chosen appropriately.
Where the gains come from
The lower bound in Theorem A is achieved by considering three cases based on the index n, leading to the final estimate. The proof relies on arguments that do not require convexity of the Schatten unit ball or quasiBanach duality, applicable even for 0 < p < 1. This approach successfully establishes the conjectured lower estimate in the range where previous results were incomplete. p. 2 Thus, the missing part of (2) concerns 0 < p < 1 and
q > 2. Our main result establishes the conjectured lower estimate for any fixed proportion of the ambient
dimension.Where the paper says it
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Improved Gelfand number estimation for Schatten class embeddings: The new theorem establishes a complete all-index description for Gelfand numbers of mappings between finite-dimensional Schatten classes in the range 0 < p <= 1 and 2 <= q <= infinity, providing sharp lower bounds that were previously missing.
-
Enhanced low-rank matrix recovery capabilities: The results provide exact asymptotic orders for approximation quantities, which can be used to design more stable and accurate algorithms for low-rank matrix recovery in settings where Schatten quasi-norms are employed.
-
Precise characterization of non-convex recovery stability: The connection between Gelfand numbers and the stability of non-convex recovery procedures by Schatten quasi-norm minimization is strengthened, allowing researchers to determine the exact limitations on stable recovery from linear measurements.
Abstract
We prove Conjecture 1.4 of Prochno and Strzelecki [J. Approx. Theory 277 (2022), 105736] on Gelfand numbers of embeddings between finite-dimensional Schatten classes. For 0<p at most1 and indices bounded away from the ambient dimension by a fixed proportion, we obtain the conjectured lower bound, uniformly in the target exponent 2 at most q at most infinity. Together with the known upper estimate obtained by Hinrichs, Prochno and Vybíral [Math. Ann., 380(3-4):1563-1593, 2021], this gives the exact asymptotic order in the previously unresolved range. As a consequence, we obtain a complete all-index description by combining our result with the estimates of Hinrichs, Prochno and Vybíral and of Prochno and Strzelecki. The proof is deterministic and is based on Hilbert--Schmidt projections on rectangular matrix spaces.
Related papers
- Local-to-Global Isometries in Standard-Norm Generalized Gyrovector Spaces
- Banach Gelfand Triples with Kernel Theorems via Equivalence Bimodules
- On the dynamics of composition operators: supercyclicity, odometers and translations
- The Local Embedding Problem for Hardy Spaces of Dirichlet Series
- The Strong Naor--Schechtman Convolution Inequality on the Product of Cyclic Groups
- Structured coordinate systems in L p(R)