Lifts of Operator Systems
Markus Dannemüller, Tim Netzer · 2025
Read Lifts of Operator Systems by Markus Dannemüller, Tim Netzer (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-2508-15348.txt.
Introduction
Markus Dannemuller and Tim Netzer carry a theory of lifts from convex cones over to operator systems, in a 2025 paper. For cones, slack operators and sums-of-squares tests judge when a set can be written as the shadow of a simpler one. The pair rebuild those tools for operator systems. Their tools address a sharp question. Given an abstract operator system, is the algebra that wraps it finite in size? Or is the system the image of one whose wrapping algebra is finite?
Catalog record
| CREATOR | Markus Dannemüller, Tim Netzer |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 30 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2508-15348
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Lifts of Operator Systems Markus Dannemüller, Tim Netzer Abstract We transfer the theory of slack operators and sums-of-squares-criteria for lifts from convex cones to operator systems. These allow to study the following question, among others: Given an abstract operator system, is its enveloping C ∗ C^{*} -algebra finite-dimensional, or is it the linear image of a system with a finite-dimensional enveloping algebra? 1 Introduction The study of convex sets and their representations through more tractable objects, such as spectrahedra or free spectrahedra, lies at the heart of modern convex optimization and real algebraic geometry. A central question is whether a given convex set can be represented as a linear image of an affine section of a cone of positive semidefinite matrices. Such representations, known as semidefinite representations or spectrahedral lifts, are highly desirable due to the availability of efficient algorithms for semidefinite programming. In the classical (commutative) setting, foundational work from [ 5 , 11 ] established a deep connection between the existence of such lifts and the factorization of slack operators. This line of research was further enriched by sum-of-squares characterizations, which link the existence of a lift to the ability to express all positive linear functionals as sums of squares with respect to a finite-dimensional function space [ 2 , 3 , 10 ] . This paper extends this framework to the noncommutative setting of operator systems. Operator systems provide a natural generalization of convex cones to the matrix-valued context, capturing the structure of positivity across all matrix levels. They play a fundamental role in operator theory and quantum information. We introduce and study noncommutative lifts and factorizations for operator systems, mirroring and generalizing the classical notions. Our main results show that an operator system admits a lift to another operator system T if and only if its (noncommutative) slack operator admits a T -factorization. This extends the main theorem from [ 5 ] to the operator system setting. We further develop a sum-of-squares characterization for lifts of minimal operator systems – those whose positivity structure is entirely determined by a base cone. This allows us to derive obstructions to the existence of free spectrahedral lifts using tools from real algebraic geometry, including the construction of positive matrix polynomials that are not sums of hermitian squares. Our work not only unifies and extends existing lift-and-factorization theories but also opens new pathways for studying finite-dimensionality and representability in operator systems. The results have implications for the structure of enveloping C^{*} -algebras and the representation of convex sets in quantum and noncommutative optimization. 2 Preliminaries 2.1 Lifts and Factorizations for Convex Cones In this section, we collect important definitions and results on slack operators, sum-of-squares certificates for positive functionals, and lifts for convex cones. In Section 3 we will transfer most of these concepts and results to the non-commutative setup of operator systems. All vector spaces in this paper are real and equipped with the finest locally convex topology, all topological notions refer to it. A convex cone C\subseteq V in a vector space V is called proper if it is pointed (i.e. C\cap-C=\{0\} ), closed and has non-empty interior. Given a convex cone C , we call a subset a generating set for C , usually denoted \chi(C) , if C=\operatorname{cc}(\chi(C)) , where \operatorname{cc}(\cdot) denotes the conic hull. For a given convex cone C\subseteq V , one can ask whether it is representable in terms of a different convex cone. This is of interest in particular when the other cone is in some sense more well-behaved than C . More precisely, given a convex cone K\subseteq W , we say that C admits a K -realization , if C\cong H\cap K for some subspace H\subseteq W . The realization is proper , if H intersects the interior of K . C is said to have a K -lift , if it is the linear image of a cone that has a K -realization. The lift is called proper if the corresponding realization is proper. A particularly interesting family of such lifts are those where K=\operatorname{Psd}_{d}(\mathbb{C}) is a cone of positive semidefinite hermitian matrices. In this case, a cone with a K -realization is also called a spectrahedron , and a cone with a K -lift a spectrahedrop . It was shown in [ 5 ] that the existence of K -lifts has different interesting characterisations. We note at this point that many of the following definitions and results are often formulated for bounded convex sets, but they can be adapted easily to the conic case. First, one defines the notion of a slack operator and of a K -factorization: Definition 2.1 . Let C be a proper convex cone with dual cone C^{\lor}\subseteq V^{\prime} , and let \chi(C) and \chi(C^{\lor}) be generating sets for C and C^{\lor} , respectively. The map \langle\cdot,\cdot\rangle_{C}\colon C\times C^{\vee}\to\mathbb{R}_{\geqslant 0};\ (c,f)\mapsto f(c) is called the slack operator of C . A pair of (possibly non-linear) maps \alpha\colon C\to K,\quad\beta\colon C^{\lor}\to K^{\lor} is called K -factorization for the slack operator of C , if for all c\in\chi(C) and for all f\in\chi(C^{\lor}) it holds that \langle c,f\rangle_{C}=\langle\alpha(c),\beta(f)\rangle_{K}. One of the main results in [ 5 ] is the following theorem, which shows that the existence of K -lifts and K -factorizations is essentially equivalent, with one direction requiring a mild additional assumption: Theorem 2.2 . Let C and K be finite-dimensional proper convex cones. If C has a proper K -lift, then its slack operator has a K -factorization. Conversely, if the slack operator of C has a K -factorization, then C has a K -lift. Since we will be mostly interested in spectrahedral lifts, the following sum-of-squares formulation of the existence of a \operatorname{Psd}_{d}(\mathbb{C}) -lift is also of interest. The sufficient condition for a lift goes back to [ 7 ] , the necessity was first shown in [ 10 ] , and slightly adapted in [ 2 ] (see also [ 3 , 8 ] for more details). Theorem 2.3 . Let C be a finite-dimensional proper convex cone and let W\subseteq\mathcal{F}(C,\mathbb{R}) be a finite-dimensional subspace of functions such that for any f\in C^{\lor} , we find a sum-of-squares representation with elements h_{k}\in W , i.e. f(c)=\sum_{k}h_{k}(c)^{2} for all c\in\chi(C) . Then C admits a \operatorname{Psd}_{d}(\mathbb{C}) -lift for d=\operatorname{dim}(W) . Conversely, if C admits a \operatorname{Psd}_{d}(\mathbb{C}) -lift, then there exists a subspace W of functions with the above property and with \operatorname{dim}(W)\leqslant d^{2} . 2.2 Operator Systems In this section we collect some basic facts about operator systems, see [ 9 ] for more detailed information and references. An abstract operator system S on the vector space X consists of a proper convex cone S_{s}\subseteq X\otimes{\rm Her}_{s}(\mathbb{C}) for each s\geqslant 1, such that for all V\in{\rm Mat}_{s,t}(\mathbb{C}) the linear map {\rm id}_{X}\otimes(V^{*}\cdot V)\colon X\otimes{\rm Her}_{s}(\mathbb{C})\to X\otimes{\rm Her}_{t}(\mathbb{C}) is positive with respect to these cones, i.e. maps S_{s} into S_{t} . Here we denote by {\rm Her}_{s}(\mathbb{C}) the vector space of complex Hermitian matrices. For each Hilbert space H we denote by \mathbb{B}_{\rm sa}(H) the space of self-adjoint bounded operators, and we use the identification \mathbb{B}_{\rm sa}(H)\otimes{\rm Her}_{s}(\mathbb{C})=\mathbb{B}_{\rm sa}(H^{s}) to define an important operator system P_{H} on the space \mathbb{B}_{\rm sa}(H) , defined as (P_{H})_{s}\coloneqq\mathbb{B}_{\rm