Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

QUANTUM DECOMPOSITION AND QUANTUM CENTRAL LIMIT THEOREM

Akihito Hora, Nobuaki Obata

Published Feb 1, 2003• 15 citations
Publisher

Abstract

On the basis of the canonical relation between an interacting Fock space and a system of orthogonal polynomials we introduce the notion of quantum decomposition of a real random variable in an algebraic probability space. To understand prototypes we review some basic examples appearing from the Boson, Fermion, and free Fock spaces. We then prove quantum central limit theorems for the quantum components of the adjacency matrices (combinatorial Laplacians) of a growing family of regular connected graphs. As a corollary, asymptotic properties of the adjacency matrix are obtained. Concrete examples include lattices, homogeneous trees, Cayley graphs of the Coxeter groups, Hamming graphs and Johnson graphs. In particular, asymptotic spectral distribution of the adjacency matrix of a Johnson graph is described by an interacting Fock space corresponding to the Meixner polynomials which are one-parameter deformation of the Laguerre polynomials.

DOI
10.1142/9789812704412_0016

Related papers

The quantum moment problem for a classical random variable and a classification of interacting Fock spaces

Luigi Accardi, Yun Gang Lu

The fact that any classical random variable with all moments has a quantum decomposition allows to associate to it a family of quantum moments. On the other hand, a classical random variable may have several inequivalent quantum decompositions, which lead to the same classical, but different quantum moments. Even in the simplest Central Limit Theorems (CLT), i.e. those of Bernoulli type, there are examples in which the corresponding quantum moments converge to the canonical quantum moments of the associated classical random variable, and examples in which this is not the case. This poses the problem to find a constructive criterium that characterizes the quantum moments associated to the canonical quantum decomposition (which is unique) with respect to the other ones. Theorem [Formula: see text] of the present paper provides such a criterium. Theorem [Formula: see text] deals with the case when one knows a priori that the quantum moments come from a central limit theorem (the motivation of the present paper arose in this context). It gives only a sufficient condition, but simpler to verify than the necessary and sufficient conditions of Theorem [Formula: see text]. Theorem [Formula: see text] naturally leads to a classification of Interacting Fock Spaces (IFS) into three types. We construct examples showing that all these possibilities can effectively take place. On the way, we prove that all the best known deformations of Heisenberg commutation relations can be obtained as special cases of a general construction within the algebraic approach to the theory of orthogonal polynomials.

Infinite Dimensional Analysis Quantum Probability and Related Topics•Mar 1, 2022•8 citations

THE QUANTUM DECOMPOSITION OF RANDOM VARIABLES WITHOUT MOMENTS

Luigi Accardi, Habib Rebei, Anis Riahi

The quantum decomposition of a classical random variable is one of the deep results of quantum probability: it shows that any classical random variable or stochastic process has a built-in non-commutative structure which is intrinsic and canonical, and not artificially put by hands. Up to now the technique to deduce the quantum decomposition has been based on the theory of interacting Fock spaces and on Jacobi's tri-diagonal relation for orthogonal polynomials. Therefore it requires the existence of moments of any order and cannot be applied to random variables without this property. The problem to find an analogue of the quantum decomposition for random variables without finite moments of any order remained open for about fifteen years and nobody had any idea of how such a decomposition could look like. In the present paper we prove that any infinitely divisible random variable has a quantum decomposition canonically associated to its Lévy–Khintchin triple. The analytical formulation of this result is based on Kolmogorov representation of these triples in terms of 1–cocycles (helices) in Hilbert spaces and on the Araki–Woods–Parthasarathy–Schmidt characterization of these representation in terms of Fock spaces. It distinguishes three classes of random variables: (i) with finite second moment; (ii) with finite first moment only; (iii) without any moment. The third class involves a new type of renormalization based on the associated Lévy–Khinchin function.

Infinite Dimensional Analysis Quantum Probability and Related Topics•Jun 1, 2013•15 citations

CLIFFORD ALGEBRAS, RANDOM GRAPHS, AND QUANTUM RANDOM VARIABLES

René Schott, G. Stacey Staples

For fixed n > 0, the space of finite graphs on n vertices is canonically associated with an abelian, nilpotent-generated subalgebra of the Clifford algebra which is canonically isomorphic to the 2n-particle fermion algebra. Using the generators of the subalgebra, an algebraic probability space of “Clifford adjacency matrices” associated with finite graphs is defined. Each Clifford adjacency matrix is a quantum random variable whose m th moment corresponds to the number of m-cycles in the graph G. Each matrix admits a canonical “quantum decomposition” into a sum of three algebraic random variables: a = a Δ + a ϒ + a Λ , where a Δ is classical while a ϒ and a Λ are quantum. Moreover, within the Clifford algebra context the NP problem of cycle enumeration is reduced to matrix multiplication, requiring no more than n 4 Clifford (geo-metric) multiplications within the algebra.

Quantum Stochastics and Information•Aug 1, 2008•1 citation

Nilpotent adjacency matrices, random graphs and quantum random variables

René Schott, G. Stacey Staples

While a number of researchers have previously investigated the relationship between graph theory and quantum probability, the current work explores a new perspective. The approach of this paper is to begin with an arbitrary graph having no previously established relationship to quantum probability and to use that graph to construct a quantum probability space in which moments of quantum random variables reveal information about the graph's structure. Given an arbitrary finite graph and arbitrary odd integer m ⩾ 3, fermion annihilation operators are used to construct a family of quantum random variables whose m th moments correspond to the graph's m -cycles. The approach is then generalized to recover a graph's m -cycles for any integer m ⩾ 3 by defining nilpotent adjacency operators in terms of null-square generators of an infinite-dimensional Abelian algebra. It is shown that ordering the vertices of a simple graph induces a canonical decomposition Ψ = Ψ + + Ψ − on any nilpotent adjacency operator Ψ. The work concludes with applications to Markov chains and random graphs.

Journal of Physics A Mathematical and Theoretical•Apr 2, 2008•18 citations

The qq-bit (III): Symmetric q-Jordan–Wigner embeddings

Luigi Accardi, Yun Gang Lu

We prove that, replacing the left Jordan–Wigner [Formula: see text]-embedding by the symmetric [Formula: see text]-embedding described in Sec. 2, the result of the corresponding central limit theorem changes drastically with respect to those obtained in Ref. 5. In fact, in the former case, for any [Formula: see text], the limit space is precisely the [Formula: see text]-mode Interacting Fock Space (IFS) that realizes the canonical quantum decomposition of the limit classical random variable. In the latter case, this happens if and only if [Formula: see text]. Furthermore, as shown in Sec. 4, the limit classical random variable turns out to coincide with the [Formula: see text]-mode version of the [Formula: see text]-deformed quantum Brownian introduced by Parthasarathy[Formula: see text], and extended to the general context of bi-algebras by Schürman[Formula: see text]. The last section of the paper (Appendix) describes this continuous version in white noise language, leading to a simplification of the original proofs, based on quantum stochastic calculus.

Infinite Dimensional Analysis Quantum Probability and Related Topics•Nov 12, 2018•1 citation

CONSTRUCTIVE UNIVERSAL CENTRAL LIMIT THEOREMS BASED ON INTERACTING FOCK SPACES

Luigi Accardi, Vitonofrio Crismale, Yun Gang Lu

Cabana-Duvillard and lonescu11have proved that any symmetric probability measure with moments of any order can be obtained as central limit theorem of self-adjoint, weakly independent and symmetrically distributed (in a quantum souse) random variables. Results of this type will be called "universal central limit theorem". Using Interacting Fock Space (IFS) techniques we extend this result in two directions: (i) we prove that the random variables can be taken to be generalized Gaussian in the sense of Accardi and Bożejko3and we give a realization of such random variables as sums of creation, annihilation and preservation operators acting on an appropriate IFS; (ii) we extend the above-mentioned result to the nonsymmetric case. The nontrivial difference between the symmetric and the nonsymmetric case is explained at the end of the introduction below.

Infinite Dimensional Analysis Quantum Probability and Related Topics•Nov 22, 2005•18 citations
On the basis of the canonical relation between an interacting Fock space and a system of orthogonal polynomials we introduce the notion of quantum decomposition of a real random variable in an algebraic probability space. To understand prototypes we review some basic examples appearing from the Boson, Fermion, and free Fock spaces. We then prove quantum central limit theorems for the quantum components of the adjacency matrices (combinatorial Laplacians) of a growing family of regular connected graphs. As a corollary, asymptotic properties of the adjacency matrix are obtained. Concrete examples include lattices, homogeneous trees, Cayley graphs of the Coxeter groups, Hamming graphs and Johnson graphs. In particular, asymptotic spectral distribution of the adjacency matrix of a Johnson graph is described by an interacting Fock space corresponding to the Meixner polynomials which are one-parameter deformation of the Laguerre polynomials.
The fact that any classical random variable with all moments has a quantum decomposition allows to associate to it a family of quantum moments. On the other hand, a classical random variable may have several inequivalent quantum decompositions, which lead to the same classical, but different quantum moments. Even in the simplest Central Limit Theorems (CLT), i.e. those of Bernoulli type, there are examples in which the corresponding quantum moments converge to the canonical quantum moments of the associated classical random variable, and examples in which this is not the case. This poses the problem to find a constructive criterium that characterizes the quantum moments associated to the canonical quantum decomposition (which is unique) with respect to the other ones. Theorem [Formula: see text] of the present paper provides such a criterium. Theorem [Formula: see text] deals with the case when one knows a priori that the quantum moments come from a central limit theorem (the motivation of the present paper arose in this context). It gives only a sufficient condition, but simpler to verify than the necessary and sufficient conditions of Theorem [Formula: see text]. Theorem [Formula: see text] naturally leads to a classification of Interacting Fock Spaces (IFS) into three types. We construct examples showing that all these possibilities can effectively take place. On the way, we prove that all the best known deformations of Heisenberg commutation relations can be obtained as special cases of a general construction within the algebraic approach to the theory of orthogonal polynomials.
The quantum decomposition of a classical random variable is one of the deep results of quantum probability: it shows that any classical random variable or stochastic process has a built-in non-commutative structure which is intrinsic and canonical, and not artificially put by hands. Up to now the technique to deduce the quantum decomposition has been based on the theory of interacting Fock spaces and on Jacobi's tri-diagonal relation for orthogonal polynomials. Therefore it requires the existence of moments of any order and cannot be applied to random variables without this property. The problem to find an analogue of the quantum decomposition for random variables without finite moments of any order remained open for about fifteen years and nobody had any idea of how such a decomposition could look like. In the present paper we prove that any infinitely divisible random variable has a quantum decomposition canonically associated to its Lévy–Khintchin triple. The analytical formulation of this result is based on Kolmogorov representation of these triples in terms of 1–cocycles (helices) in Hilbert spaces and on the Araki–Woods–Parthasarathy–Schmidt characterization of these representation in terms of Fock spaces. It distinguishes three classes of random variables: (i) with finite second moment; (ii) with finite first moment only; (iii) without any moment. The third class involves a new type of renormalization based on the associated Lévy–Khinchin function.
For fixed n > 0, the space of finite graphs on n vertices is canonically associated with an abelian, nilpotent-generated subalgebra of the Clifford algebra which is canonically isomorphic to the 2n-particle fermion algebra. Using the generators of the subalgebra, an algebraic probability space of “Clifford adjacency matrices” associated with finite graphs is defined. Each Clifford adjacency matrix is a quantum random variable whose m th moment corresponds to the number of m-cycles in the graph G. Each matrix admits a canonical “quantum decomposition” into a sum of three algebraic random variables: a = a Δ + a ϒ + a Λ , where a Δ is classical while a ϒ and a Λ are quantum. Moreover, within the Clifford algebra context the NP problem of cycle enumeration is reduced to matrix multiplication, requiring no more than n 4 Clifford (geo-metric) multiplications within the algebra.
While a number of researchers have previously investigated the relationship between graph theory and quantum probability, the current work explores a new perspective. The approach of this paper is to begin with an arbitrary graph having no previously established relationship to quantum probability and to use that graph to construct a quantum probability space in which moments of quantum random variables reveal information about the graph's structure. Given an arbitrary finite graph and arbitrary odd integer m ⩾ 3, fermion annihilation operators are used to construct a family of quantum random variables whose m th moments correspond to the graph's m -cycles. The approach is then generalized to recover a graph's m -cycles for any integer m ⩾ 3 by defining nilpotent adjacency operators in terms of null-square generators of an infinite-dimensional Abelian algebra. It is shown that ordering the vertices of a simple graph induces a canonical decomposition Ψ = Ψ + + Ψ − on any nilpotent adjacency operator Ψ. The work concludes with applications to Markov chains and random graphs.
We prove that, replacing the left Jordan–Wigner [Formula: see text]-embedding by the symmetric [Formula: see text]-embedding described in Sec. 2, the result of the corresponding central limit theorem changes drastically with respect to those obtained in Ref. 5. In fact, in the former case, for any [Formula: see text], the limit space is precisely the [Formula: see text]-mode Interacting Fock Space (IFS) that realizes the canonical quantum decomposition of the limit classical random variable. In the latter case, this happens if and only if [Formula: see text]. Furthermore, as shown in Sec. 4, the limit classical random variable turns out to coincide with the [Formula: see text]-mode version of the [Formula: see text]-deformed quantum Brownian introduced by Parthasarathy[Formula: see text], and extended to the general context of bi-algebras by Schürman[Formula: see text]. The last section of the paper (Appendix) describes this continuous version in white noise language, leading to a simplification of the original proofs, based on quantum stochastic calculus.
Cabana-Duvillard and lonescu11have proved that any symmetric probability measure with moments of any order can be obtained as central limit theorem of self-adjoint, weakly independent and symmetrically distributed (in a quantum souse) random variables. Results of this type will be called "universal central limit theorem". Using Interacting Fock Space (IFS) techniques we extend this result in two directions: (i) we prove that the random variables can be taken to be generalized Gaussian in the sense of Accardi and Bożejko3and we give a realization of such random variables as sums of creation, annihilation and preservation operators acting on an appropriate IFS; (ii) we extend the above-mentioned result to the nonsymmetric case. The nontrivial difference between the symmetric and the nonsymmetric case is explained at the end of the introduction below.