Isometric Embeddability of Schatten Classes Revisited

This paper summarizes existing results and open questions regarding isometric embeddings between Schatten classes, provides an overview of relevant methods, and establishes a new non-embeddability result using a novel approach.

The Gist

Imagine you have a collection of different-sized boxes. Some are made of wood, some of glass, some of steel. In the world of mathematics, these "boxes" are called Schatten classes. They are special containers used to hold "operators" (which you can think of as complex machines that transform data).

The big question this paper asks is: Can you fit one of these boxes perfectly inside another without squishing or stretching it?

In math terms, this is called an isometric embedding. "Isometric" means "same measure." If you put a box inside another, the "size" (or norm) of everything inside must stay exactly the same. If you have to stretch a rubber band to make it fit, that's not allowed.

Here is the breakdown of the paper using simple analogies:

1. The Players: The Boxes

The authors are studying different types of boxes, labeled by a number $p$ (like $S_1, S_2, S_3$, etc.).

• The "Diagonal" Boxes: Inside these complex boxes, there are simpler, one-dimensional strips (called sequence spaces, like $\ell_p$). Think of these as the "spine" of the box. If you can't fit the spine of Box A into Box B, you definitely can't fit the whole Box A.
• The "Function" Boxes: There are also standard boxes made of continuous functions (like $L_p$ spaces). The paper discovers a secret tunnel connecting the complex Schatten boxes to these simpler function boxes.

2. The Main Discovery: The "No-Go" Zones

For a long time, mathematicians knew that some boxes could fit inside others. For example, a small square box can fit inside a larger square box.

• The Good News: If the boxes are the same "material" (same $p$ value), a smaller one fits into a larger one.
• The Bad News (The Paper's Focus): What happens if you try to fit a "wooden" box ($p=1$) inside a "glass" box ($p=2$)? Or a "steel" box ($p=\infty$) inside a "plastic" one ($p=3$)?

The authors summarize what we already know and add a new, powerful rule that says "No" to many more combinations than before.

3. The New Tool: The "X-Ray Machine"

The authors introduce a novel method to prove these boxes don't fit.

• Old Method: They used to try to measure the boxes directly, but for some tricky materials (where $p < 2$), the "ruler" broke. The math got too jagged to measure with standard tools.
• The New Method: They used a "magic X-ray" (based on something called multilinear operator integrals and a theorem by Kato-Rellich).
  • Imagine you have a machine that vibrates. If you shake Box A and Box B, the way they vibrate tells you their internal structure.
  • The authors found that if you try to force a Box A into Box B, the "vibrations" (mathematical derivatives) don't match up. The math screams, "This doesn't fit!" because the shape of the vibration is fundamentally different.

4. The "Bridge" Analogy

The most exciting part of the paper is a new bridge they built between two worlds:

1. The Quantum World: The complex, non-commutative Schatten classes (where order matters: $A \times B \neq B \times A$).
2. The Classical World: Simple, everyday function spaces (where order doesn't matter).

The Analogy:
Imagine you are trying to fit a 3D puzzle piece (Schatten class) into a 2D puzzle (Function space).

• The authors proved that if you can fit the 3D piece into the 2D puzzle, then a specific "shadow" of that piece (a sequence space) must also fit into the 2D puzzle.
• But, they already knew from previous work that this specific shadow cannot fit into that 2D puzzle.
• Conclusion: Therefore, the 3D piece cannot fit either.

This allowed them to prove that for many combinations of $p$ and $q$ (specifically when $q < p$ and $p > 1$), it is impossible to embed one Schatten class into another isometrically.

5. What's Still a Mystery? (The Open Problems)

Even with their new tools, some boxes remain a mystery. The authors list the "unsolved cases" like a detective board:

• The "Square" Mystery: Can a 2D square ($S_2$) fit into a box of size $p$ if the box is smaller than a certain limit? We don't know yet.
• The "Infinity" Mystery: Can a box made of infinite materials fit into a box made of finite materials in specific weird scenarios?
• The "Small $p$" Mystery: When the numbers get very small (less than 1), the math gets very weird, and they haven't figured out all the rules yet.

Summary

This paper is a "status report" on fitting mathematical boxes into each other.

1. Recap: It lists all the rules we already knew.
2. New Proof: It uses a clever new "vibration test" (multilinear integrals) and a "bridge" to classical spaces to prove that many more combinations are impossible than we thought.
3. Future Work: It points out the remaining "impossible" cases that are actually still possible, inviting other mathematicians to solve the next puzzle.

In short: You can't just stuff any mathematical box into any other box. The authors proved exactly which combinations are impossible, using a new method that looks at how these boxes "vibrate" when you try to force them together.

Technical Summary

Here is a detailed technical summary of the paper "Isometric Embeddability of Schatten Classes Revisited" by Chattopadhyay, Pradhan, and Skripka.

1. Problem Statement

The central problem addressed in this note is the existence of linear isometric embeddings between different Schatten $p$-classes, denoted as $S_p(H)$, where $H$ is a complex separable Hilbert space.

Specifically, the authors investigate Problem 1.1, which asks under what conditions an isometric embedding exists from $S_q^m$ (finite-dimensional Schatten class of dimension $m$) or $S_q(H)$ (infinite-dimensional) into $S_p^n$ or $S_p(H)$, respectively, given $p \neq q$.

The problem is contextualized within the broader study of:

• Sequence spaces: $\ell_p^n(K)$ (where $K$ is $\mathbb{R}, \mathbb{C},$ or $\mathbb{H}$).
• Function spaces: $L_p(\Omega, d\mu)$.
• Commutative subspaces: The authors note that commutative subspaces of Schatten classes are isometrically isomorphic to sequence spaces $\ell_p$, making the embeddability of $\ell_p$ into $S_p$ a foundational step.

2. Methodology

The paper synthesizes classical results with novel techniques to establish both affirmative and negative embeddability results. The methodologies employed include:

• Classical Isometry Characterization: Utilizing Banach's and Lamperti's characterizations of isometries on $\ell_p$ and $L_p$ spaces to establish baseline non-embeddability results.
• Multilinear Operator Integrals: A core technique in the paper involves the use of multilinear operator integrals (specifically the discrete multilinear operator integral $T_{f[n]}^{A_0, \dots, A_n}$). This allows the authors to compute derivatives of the Schatten $p$-norm with respect to perturbations.
  • They utilize the formula for the second derivative of the norm: $\frac{d^2}{dt^2}\|A + tB\|_p^p |_{t=0} = 2! \text{Tr}(T_{f[2]}^{A,A,A}(B, B))$, where $f(x) = |x|^p$.
• Analytic Perturbation Theory: The Kato–Rellich theorem is applied to assert that the eigenvalues of a linear perturbation of a diagonal matrix are locally analytic functions of the perturbation parameter. This allows for a comparison of the analytic behavior of the norm on both sides of a hypothetical embedding equation.
• Reduction to Function Spaces: For cases where $p < 2$ (where the second derivative of the norm does not exist), the authors employ a novel reduction strategy. They use a recent result by Heinävaara [12] to show that the span of two operators in $S_p(H)$ is isometric to a subspace of a classical function space $L_p([0, 1], d\mu)$. This bridges the gap between non-commutative and commutative settings.
• Contradiction via Analytic Behavior: The primary proof strategy for non-embeddability involves assuming an isometric embedding exists, constructing a specific 2-dimensional subspace (isomorphic to $\ell_2^q$), and deriving a contradiction by comparing the analytic properties (differentiability at $t=0$) of the resulting norm equations.

3. Key Contributions and Results

A. Affirmative Embeddings (Existence)

The paper confirms several known embeddings:

• Monotonicity: $\ell_p^m \hookrightarrow \ell_p^n \hookrightarrow \ell_p$ and $S_p^m \hookrightarrow S_p^n \hookrightarrow S_p(H)$ for all $p \in (0, \infty]$.
• Special Cases:
  • $\ell_1^2(\mathbb{R}) \hookrightarrow \ell_\infty^n(\mathbb{R})$.
  • $\ell_2^m(K) \hookrightarrow \ell_{2p}^n(K)$ for $p \in \mathbb{N}$.
  • Crucial Result: $S_2^m \hookrightarrow \ell_2^{m^2}(\mathbb{C}) \hookrightarrow S_p^{m^2}$ for all $p \in (0, \infty]$. This establishes that the Hilbert-Schmidt class ($p=2$) can be embedded into $S_p$ spaces of higher dimension.

B. Negative Embeddings (Non-Existence)

The authors provide a comprehensive survey of non-embeddability results, extending previous work by [4, 5, 11, 31]:

• Finite Dimensions ($S_q^m \hookrightarrow S_p^n$):
  • Non-embeddability holds for a wide range of $(q, p)$ pairs, including $(1, \infty] \setminus \{2\} \times (1, \infty)$ and $(0, 2) \setminus \{1\} \times (0, 1)$.
  • Specifically, $S_2^m \not\hookrightarrow S_p^n$ if $p$ is an odd integer ($p \in 2\mathbb{N}+1$).
• Infinite Dimensions ($S_q(H) \hookrightarrow S_p(H)$):
  • Theorem 2.11: Establishes non-embeddability for $(q, p) \in ((1, \infty)\setminus\{2\} \times [2, \infty)) \cup ([4, \infty) \times \{1\}) \cup (\{\infty\} \times (1, \infty)) \cup ((2, \infty) \times \{\infty\})$.
  • Theorem 2.14 (New Result): The authors prove that $S_q(H) \not\hookrightarrow S_p(H)$ for $q < p$ where $(q, p) \in [1, 2) \times (1, \infty)$ and for $(q, p) \in (2, \infty) \times (1, \infty)$.
  • Significance of Theorem 2.14: This result is derived using the reduction to $L_p$ function spaces (Theorem 2.13) and a known non-embeddability result for sequence spaces into function spaces (Theorem 2.6). This provides an alternative proof to the multilinear operator integral method, avoiding the need for second derivatives of the norm.

C. Open Problems

The paper explicitly lists unresolved cases in Problem 3.1, including:

1. Whether $S_2^m$ embeds into $S_p^n$ when $n < m^2$.
2. Whether $S_3^m$ embeds into $S_1^n$ or $S_\infty^n$.
3. Whether $S_1^m$ or $S_\infty^m$ embeds into $S_p^n$ for $p = 1/k$ ($k \in \mathbb{N}$).
4. Whether $S_q^m$ embeds into $S_p^n$ for $(q, p) \in (2, \infty) \times (0, 1)$.
5. Infinite-dimensional cases for $(q, p) \in (0, \infty] \times (0, 1)$ and specific ranges involving $q > p$ in $(1, 2)$.

4. Significance

• Unification of Methods: The paper successfully unifies classical Banach space theory with modern non-commutative analysis. It demonstrates how techniques from multilinear operator integration (originally developed for perturbation theory) can solve geometric problems in Banach space theory.
• Novel Non-Embeddability Proof: The derivation of Theorem 2.14 is a significant contribution. By linking the non-commutative $S_p$ spaces to commutative $L_p$ spaces via the span of two operators, the authors bypass the analytical difficulties associated with $p < 2$ (where the norm is not twice differentiable).
• Comprehensive Survey: The paper serves as a definitive reference for the current state of the art regarding isometric embeddings in Schatten classes, clearly delineating what is known, what is proven via specific methods, and what remains open.
• Clarification of Commutative vs. Non-Commutative: It highlights the subtle differences between embedding sequence spaces ($\ell_p$) and Schatten classes ($S_p$), noting that while $\ell_1^2(\mathbb{R}) \hookrightarrow \ell_\infty^n(\mathbb{R})$, the analogous embedding $\ell_1^2(\mathbb{C}) \hookrightarrow S_\infty^n$ fails.

In summary, this note advances the understanding of the geometry of non-commutative $L_p$ spaces by leveraging both deep analytic tools (multilinear integrals, Kato-Rellich) and structural reductions to classical function spaces, while providing a clear roadmap for future research in the field.