Graph Structured Operator Inequalities and Tsirelson-Type Bounds

James Tian · 2025

Read Graph Structured Operator Inequalities and Tsirelson-Type Bounds by James Tian (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2511-01525.txt.

Introduction

James Tian proves bounds on the size of combined quantum-style systems built from pairs of self-adjoint contractions, operators that never stretch a vector. The inequalities generalize the analytic structure behind the Tsirelson and CHSH bounds, famous limits on how strongly separated quantum systems can be correlated.

The estimates are dimension-free, meaning they do not grow with the size of the system, and they are written through norms of commutators and anticommutators. A graph-based version handles sparse interaction patterns, where each system touches only a few others, with constants depending only on how connected the graph is. The results connect operator theory with Bell correlations and network nonlocality, offering closed-form estimates alongside numerical semidefinite methods.

Catalog record

CREATORJames Tian
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE33 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2511-01525

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Graph Structured Operator Inequalities and Tsirelson-Type Bounds James Tian Abstract . We establish operator norm bounds for bipartite tensor sums of self-adjoint contractions. The inequalities generalize the analytic structure underlying the Tsirelson and CHSH bounds, giving dimension-free estimates expressed through commutator and anticommutator norms. A graph based formulation captures sparse interaction patterns via constants depending only on graph connectivity. The results link analytic operator inequalities with quantum information settings such as Bell correlations and network nonlocality, offering closed-form estimates that complement semidefinite and numerical methods. 1. Introduction We study universal operator norm bounds for bipartite tensor sums B=\sum^{m}_{i=1}x_{i}\otimes y_{i}, where each x_{i} and y_{i} is a self-adjoint contraction. The starting point is the simple identity underlying the CHSH and Tsirelson bounds: for self-adjoint unitaries A_{0},A_{1},B_{0},B_{1} , \mathcal{B}:=A_{0}B_{0}+A_{0}B_{1}+A_{1}B_{0}-A_{1}B_{1} satisfies \mathcal{B}^{2}=4I-\left[A_{0},A_{1}\right]\left[B_{0},B_{1}\right] , which yields the Tsirelson bound 2\sqrt{2} once commutator norms are estimated [ CHSH69 ] . This identity, first noted by Cirel’son (Tsirelson), remains central in the modern analysis of Bell inequalities and quantum correlations [ Cir80 , Kit85 , WW01 , HHHH09 , Bra11 ] . The same commutator/anticommutator expansion has intrinsic operator-theoretic value. Sharp inequalities for commutators and anticommutators in unitarily invariant norms have been studied extensively, from classical matrix analysis to recent refinements. Examples include bounds by Bhatia and Kittaneh, the Böttcher-Wenzel inequality and its variants, and work on commutator estimates for normal or positive operators [ BK98 , BW08 ] . The proofs developed here draw directly on this analytic approach but remain fully dimension-free. Operator space and noncommutative probability techniques also connect norm control of mixed products to the structure of quantum correlations. Results such as the operator-space Grothendieck theorem, noncommutative Khintchine inequalities, and analyses of XOR games show how tensor norms and commutators govern achievable correlation strengths [ PS02 , LPP91 , Pis03 ] . These perspectives link operator inequalities with the geometry of quantum information theory. In quantum information applications, Bell inequalities now extend far beyond the two-setting CHSH case. Device independent protocols, semidefinite-programming hierarchies, and network or many-body generalizations provide powerful but often computationally heavy tools. The inequalities presented here offer a complementary, closed-form approach: graph-sensitive, dimension-free estimates that directly quantify the role of commutators and anticommutators. They connect naturally with ideas underlying the NPA hierarchy, device-independent quantum key distribution, and graph based Bell tests. [ NPA07 , BCP+14 , WCD08 , Fri12 ] . A second motivation arises from structured interactions. In spin system and quantum network models, operators such as \sigma_{x}\otimes\sigma_{x}+\sigma_{y}\otimes\sigma_{y}+\sigma_{z}\otimes\sigma_{z} represent Heisenberg type couplings, while general sums \sum_{i}x_{i}\otimes y_{i} appear in lattice or graph coupled systems. The graph based inequalities developed here provide explicit norm estimates that scale with the interaction pattern via a simple combinatorial parameter, analogous to the role of Lieb-Robinson bounds in locality analysis [ LR72 , BHV06 , NS10 ] . Concretely, we establish complete graph bounds of the form \left\|B\right\|^{2}\leq m+\frac{1}{2}\sum_{i<j}\left(\left\|\left[x_{i},x_{j}\right]\right\|\left\|\left[y_{i},y_{j}\right]\right\|+\left\|\left\{x_{i},x_{j}\right\}\right\|\left\|\left\{y_{i},y_{j}\right\}\right\|\right), with equality for anticommuting Clifford families. The argument extends directly to weighted sums \left\|B_{c}\right\|^{2}\leq\sum_{i}c^{2}_{i}+\sum_{i<j}\left|c_{i}c_{j}\right|\phi_{ij}, and to sparse graphs satisfying an edge domination condition that controls non-edge interactions by averages over neighboring edges. The graph based inequalities form the second major theme of this work. They show that the operator norm of B can be bounded by the “local” commutation structure encoded in a graph G=\left(V,E\right) , with a constant C\left(G\right)=\frac{2\left(m-1\right)}{\delta}-1 depending only on the minimum degree \delta . This gives a quantitative link between operator-norm growth and graph connectivity: dense graphs recover the complete graph constant C\left(G\right)=1 , while sparse graphs yield controlled relaxations. The same reasoning extends to the weighted setting, where each term carries a scalar amplitude c_{i} , producing a unified framework that interpolates between universal and graph local inequalities. These graph based results show how combinatorial sparsity constrains noncommutative norm growth. They align with recent developments in graph-theoretic quantum correlations, network Bell inequalities, and the study of commuting graph structures in operator algebras [ Fri12 , RBB+19 , HH96 ] . They also parallel the role of graph Laplacians and adjacency operators in noncommutative harmonic analysis and matrix-valued inequalities [ Bha97 , Ver18 ] . The resulting inequalities yield two types of consequences. First, they provide explicit, analytic upper bounds on bipartite correlators \sup_{\rho}\left|tr\left(\rho B_{c}\right)\right| in terms of noncommutativity, matching the Tsirelson value in the Clifford case. Second, they yield graph dependent bounds that show how the degree structure limits or amplifies collective correlations, allowing one to infer the presence of many substantially noncommuting pairs from an observed Bell value. This directly connects to current work on graph based nonlocality and self-testing schemes using anticommuting observables [ Kan16 , MY04 ] . Our results thus complement existing semidefinite and numerical frameworks by providing simple, verifiable operator inequalities that (i) scale transparently with the number of settings, (ii) make commutator and anticommutator dependence explicit, and (iii) adapt naturally to weighted or sparse architectures. They also offer quick analytical estimates for tensor-sum operators in quantum information and operator theory, where numerical optimization may be unnecessary or infeasible. For background we refer to standard texts in quantum information, operator theory, and operator spaces, see e.g., [ Pau02 , Wat18 ] . 2. Complete and Sparse Graph Operator Bounds This section develops the main operator norm inequalities for tensor sums of self-adjoint contractions. We begin with the complete graph case, where every pair of indices interacts, leading to a universal bound that depends only on the pairwise commutator and anticommutator norms. This gives a global inequality valid for all finite families of self-adjoint contractions. We then derive a sparse graph extension, in which the interactions are restricted to the edges of a fixed graph. Under a natural domination condition on the off-edge terms, the same operator norm formulation yields a controlled estimate whose scaling depends on the minimum degree of the graph. Theorem 2.1 . Let x_{1},\dots,x_{m}\in B\left(H\right) and y_{1},\dots,y_{m}\in B\left(K\right) be self-adjoint contractions. Set B=\sum^{m}_{i=1}x_{i}\otimes y_{i}. Then (2.1) \left\|B\right\|^{2}\leq m+\frac{1}{2}\sum_{i<j}\left(\left\|\left[x_{i},x_{j}\right]\right\|\left\|\left[y_{i},y_{j}\right]\right\|+\left\|\left\{x_{i},x_{j}\right\}\right\|\left\|\left\{y_{i},y_{j}\right\}\right\|\right). In particular, if for each pair \left(i,j\right) either \left\{x_{i},x_{j}\right\}=0 or \left\{y_{i},y_{j}\right\}=0 , then (2.2) \left\|B\right\|^{2}\leq m+\frac{1}{2}\sum_{i<j}\left\|\left[x_{i},x_{j}\right]\right\|\left\|\left[y_{i},y_{j}\right]\right\|. Proof. Expand B^{2}=\sum^{m}_{i=1}x^{2}_{i}\otimes y^{2}_{i}+\sum_{i<j}\left(x_{i}x_{j}\otimes y_{i}y_{j}+x_{j}x_{i}\otimes y_{j}y_{i}\right). Define D:=\sum^{m}_{i=1}x^{2}_{i}\otimes y^{2}_{i},\qquad T_{ij}:=x_{i}x_{j}\otimes y_{i}y_{j}+x_{j}x_{i}\otimes y_{j}y_{i}\quad\left(i<j\right). Then B^{2}=D+\sum_{i<j}T_{ij} . Each x_{i},y_{i} is a self-adjoint contraction, so 0\leq x^{2}_{i}\leq I and 0\leq y^{2}_{i}\leq I in the Loewner order. Hence 0\leq x^{2}_{i}\otimes y^{2}_{i}\leq I and therefore D\leq mI. Next, write T_{ij}=\frac{1}{2}\left(\left\{x_{i},x_{j}\right\}\otimes\left\{y_{i},y_{j}\right\}+\left[x_{i},x_{j}\right]\otimes\left[y_{i},y_{j}\right]\right). The anticommutators \left\{x_{i},x_{j}\right\} and \left\{y_{i},y_{j}\right\} are self-adjoint, so \left\{x_{i},x_{j}\right\}\otimes\left\{y_{i},y_{j}\right\} is self-adjoint. The commutators \left[x_{i},x_{j}\right] and \left[y_{i},y_{j}\right] are skew-adjoint, and hence \left[x_{i},x_{j}\right]\otimes\left[y_{i},y_{j}\right] is also self-adjoint. Thus T_{ij} is self-adjoint. For any self-adjoint operator A , we have -\left\|A\right\|I\leq A\leq\left\|A\right\|I . Using this, we get \displaystyle\left\{x_{i},x_{j}\right\}\otimes\left\{y_{i},y_{j}\right\} \displaystyle\leq\left\|\left\{x_{i},x_{j}\right\}\right\|\left\|\left\{y_{i},y_{j}\right\}\right\|I, \displaystyle\left[x_{i},x_{j}\right]\otimes\left[y_{i},y_{j}\right] \displaystyle\leq\left\|\left[x_{i},x_{j}\right]\right\|\left\|\left[y_{i},y_{j}\right]\right\|I. Therefore, T_{ij}\leq\frac{1}{2}\left(\left\|\left\{x_{i},x_{j}\right\}\right\|\left\|\left\{y_{i},y_{j}\right\}\right\|+\left\|\left[x_{i},x_{j}\right]\right\|\left\|\left[y_{i},y_{j}\right]\right\|\right)I. Finally, \displaystyle B^{2} \displaystyle=D+\sum_{i<j}T_{ij} \displaystyle\leq mI+\frac{1}{2}\sum_{i<j}\left(\left\|\left\{x_{i},x_{j}\right\}\right\|\left\|\left\{y_{i},y_{j}\right\}\right\|+\left\|\left[x_{i},x_{j}\right]\right\|\left\|\left[y_{i},y_{j}\right]\right\|\right)I. Taking norms and using \left\|B\right\|^{2}=\left\|B^{2}\right\| (since B^{*}=B ) gives the claimed inequality ( 2.1 ). The assertion ( 2.2 ) follows by omitting the anticommutator terms. ∎ The complete graph bound in Theorem 2.1 treats the fully coupled case, in which every pair contributes to the mixed-term expansion of B^{2} . In many structured settings, however, only a subset of pairs interact. For example, when the operators x_{i} and y_{i} are coupled according to a network or sparsity pattern. To capture such partial coupling, we introduce a graph based formulation. The next result shows that, under an edge domination hypothesis relating non-edge to edge interactions, the same

Read the complete text