AKLT State is Indeed the Observation Process of a causal Hidden quantum Markov Model

2026

Read AKLT State is Indeed the Observation Process of a causal Hidden quantum Markov Model (2026) 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-2605-24431.txt.

Introduction

Souissi and Andolsi give a rigorous formulation of the spin-1 AKLT state inside hidden quantum Markov models. They show the AKLT ground state naturally appears as the observable output of a causal hidden quantum Markov model. That gives it hidden quantum memory in line with the standard matrix-product-state picture. This view yields a compact picture of the AKLT chain as a spin system with intrinsic quantum memory. The results suggest the hidden quantum Markov model frame suits measurement-based quantum computation and related tasks.

Catalog record

YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE64450
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.24431

Browse the full library · About the Tower of Babel

Full text

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

AKLT State is Indeed the Observation Process of a causal Hidden quantum Markov Model Abstract We present a rigorous formulation of the spin-1 Affleck–Kennedy–Lieb–Tasaki (AKLT) state within the framework of hidden quantum Markov models (HQMMs). We demonstrate that the AKLT ground state admits a natural representation as the observable output of a causal HQMM, thereby endowing it with an underlying hidden quantum memory consistent with its standard finitely correlated (matrix product state) description. This perspective provides a compact and structurally transparent characterization of the AKLT chain as a quantum spin system equipped with intrinsic quantum memory. Our results suggest that the HQMM perspective may offer a useful framework for investigating measurement-based quantum computation(MBQC). Abdessatar Souissi Department of Management Information Systems, College of Business and Economics, Qassim University, Buraydah 51452, Saudi Arabia [email protected] Amenallah Andolsi Nuclear Physics and High Energy Research Unit, Faculty of Sciences of Tunis, Tunis El Manar University, Tunis, Tunisia [email protected] Keywords : AKLT State, Hidden quantum Markov Models, Matrix Product States, hidden memory 1 Introduction The AKLT model provides a canonical and analytically tractable example of a one-dimensional quantum spin system in a symmetry-protected topological (SPT) phase [ 1 , 2 , 3 , 4 , 5 ] . It is an exactly solvable, translation-invariant spin-1 antiferromagnetic chain (with local Hilbert space \mathcal{H}_{1}\equiv\mathbb{C}^{3} ) in the Haldane phase [ 6 ] , whose unique ground state is gapped, exhibits exponentially decaying correlations, and supports fractionalized spin- \tfrac{1}{2} edge modes associated with a virtual Hilbert space \mathcal{H}_{1/2}\equiv\mathbb{C}^{2} [ 7 , 8 ] .Beyond its structural role in quantum many-body theory, the AKLT state has gained renewed significance as a resource for measurement-based quantum computation (MBQC), where its structured entanglement and hidden nonlocal (string) order enable universal quantum computation and quantum communication protocols [ 9 , 10 , 11 , 12 , 13 , 14 ] . In particular, this hidden order supports the processing of quantum information through adaptive measurement sequences that induce effective dynamical transformations on an underlying virtual space [ 15 , 16 ] . Furthermore, the ground state of the AKLT model admits a natural description both as a finitely correlated state (FCS) and as a matrix product state (MPS) [ 17 , 18 , 19 ] . Concretely, for a finite chain of n spins, the AKLT ground state can be written as the MPS |\psi_{\mathrm{AKLT}}^{(n)}\rangle=\sum_{k_{1},\dots,k_{n}\in\{+,0,-\}}\operatorname{Tr}(A_{k_{1}}\cdots A_{k_{n}})\,|k_{1}\cdots k_{n}\rangle, where the matrices A_{k} implement the local virtual-to-physical mapping. The expectation value of a local observable Y is given by \omega_{n}(Y)=\langle\psi_{\mathrm{AKLT}}^{(n)}|Y|\psi_{\mathrm{AKLT}}^{(n)}\rangle. In the thermodynamic limit, the finitely correlated state construction [ 17 ] yields \omega(Y):=\lim_{\begin{subarray}{c}m\to\infty\\ p\to\infty\end{subarray}}\omega_{n+m+p}\bigl(\mathbb{I}^{\otimes m}\otimes Y\otimes\mathbb{I}^{\otimes p}\bigr), which defines a translation-invariant state on the quasi-local algebra \mathcal{B}(\mathcal{H}_{1})^{\otimes\mathbb{N}} . Its correlation functions are fully determined by the spectral properties of the associated transfer operator, making the AKLT model a natural framework for operator-algebraic approaches to topological quantum phases and hidden quantum Markov structures. HQMMs have attracted significant attention over the past decade, motivated by their ability to capture non-classical correlations and memory effects inherent to quantum systems [ 20 , 21 , 22 ] . In particular, recent advances in the implementation, simulation, and training of HQMM have established them as promising architectures for quantum machine learning, capable of encoding complex temporal correlations that remain inaccessible to classical stochastic models [ 34 , 26 , 28 ] . Beyond their algorithmic relevance, HQMM provide a rigorous operational framework for investigating the structure of quantum memory, its accessibility through measurement, and the fundamental constraints governing its expressive power [ 29 , 32 ] , revealing deep connections between memory and fundamental physical resources such as compression efficiency, thermodynamic cost, and the emergence of adaptive behavior in quantum agents [ 33 ] . Building on these developments, HQMM-based paradigms are increasingly integrated into broader learning frameworks, including reinforcement learning for quantum processes with hidden memory [ 30 ] . A natural hidden Markov model structure on the AKLT chain is captured by an HQMM, which is a quantum stochastic process on (\mathcal{B}(\mathcal{H}_{\frac{1}{2}})\otimes\mathcal{B}(\mathcal{H}_{1}))^{\otimes\mathbb{N}} , generated by a completely positive unital map \mathcal{T}:\mathcal{B}(\mathcal{H}_{\frac{1}{2}})\otimes\mathcal{B}(\mathcal{H}_{\frac{1}{2}})\otimes\mathcal{B}(\mathcal{H}_{1})\longrightarrow\mathcal{B}(\mathcal{H}_{\frac{1}{2}}), where the underlying process is a quantum Markov chain that generates the hidden memory of the system [ 23 , 24 ] . From this general three-tensor map, two particular causal orderings emerge as special cases: the conventional ordering \mathcal{T}^{(\rm conv)}(a\otimes b\otimes x)=\mathcal{E}_{H}\bigl(\mathcal{E}_{H,O}(a\otimes b)\otimes x\bigr) and the causal ordering \mathcal{T}^{(\rm caus)}(a\otimes b\otimes x)=\mathcal{E}_{H,O}\bigl(\mathcal{E}_{H}(a\otimes x)\otimes b\bigr) for a hidden transition \mathcal{E}_{H}:\mathcal{B}(\mathcal{H}_{\frac{1}{2}})\otimes\mathcal{B}(\mathcal{H}_{\frac{1}{2}})\rightarrow\mathcal{B}(\mathcal{H}_{\frac{1}{2}}) and an emission expectation \mathcal{E}_{H,O}:\mathcal{B}(\mathcal{H}_{\frac{1}{2}})\otimes\mathcal{B}(\mathcal{H}_{1})\rightarrow\mathcal{B}(\mathcal{H}_{\frac{1}{2}}) , which yield distinct information flows and correlation structures [ 27 ] . When restricted to commutative subalgebras, HQMMs reduce to classical hidden Markov models [ 25 ] , forming a unified framework for classical and quantum hidden memory processes. Our recent work [ 19 ] has successfully reformulated the AKLT state within a conventional hidden quantum Markov model (HQMM) framework, thereby translating its symmetry-protected topological (SPT) order and entanglement structure into the language of quantum probability. While this yields a natural HQMM description of the AKLT chain, it simultaneously demonstrates that the AKLT state nevertheless resists realization as an observation process—a result that aligns with the findings obtained for the entangled hidden Markov model (EHMM) [ 35 ] within the original entangled-hidden-Markov formalism [ 34 ] . Motivated by this obstruction, subsequent constructions treat the physical spin‑1 chain as an observation layer of a hidden spin‑ \tfrac{1}{2} system driven by quantum channels closely related to the AKLT transfer matrix, with SPT order encoded as a covariance property of the emission maps under projective symmetry actions. In parallel, a causal HQMM formalism has been developed in [ 27 ] , in which the causal order between hidden transitions and emission maps is reversed, leading to a genuinely different class of HQMMs compared with the previously introduced “conventional” models [ 20 ] . This enlarged landscape of HQMM architectures opens the possibility of associating hidden quantum memory representations to a broader class of many-body states. This brings us to the central question of the present work: Can the canonical finitely correlated/MPS representation of the AKLT ground state be realised as the observation process of a suitably defined HQMM? We begin by constructing the AKLT ground state as an infinite-volume FCS on the observable algebra [ 17 ] . This FCS representation naturally allows us to express the model rigorously in the Heisenberg picture, yielding a translation-invariant state \omega fully characterized by an explicit transfer-operator formula for arbitrary local observables. Building on this algebraic foundation, we introduce a class of causal HQMM and establish that their joint hidden–observable states admit a Kraus-type decomposition, thereby making explicit how hidden quantum memory and emission mechanisms cooperate to generate correlations along the chain. As a consequence, we prove that the AKLT ground state \omega arises precisely as the observation process \psi_{O}=\psi_{H,O}\mid_{\mathcal{B}(\mathcal{H}_{1})^{\otimes\mathbb{N}}} of a causal HQMM. In sharp contrast, such a realization is provably impossible within the conventional HQMM framework applied to the AKLT state [ 19 ] , and similarly fails within the EHMMs setting [ 35 ] . This striking separation between causal and conventional HQMM representations reveals a deeper structural distinguishability between the two causal architectures—one that reflects the interplay between quantum hidden memory, measurement ordering, and topological quantum phases—and will be the subject

Read the complete text