Unification of Independence in Quantum Probability

From MaRDI portal



Abstract: For a family of unital free *-algebras with a family of states on them, we construct a sequence of noncommutative probability spaces, which are tensor product algebras with tensor product states and which approximate the free product of states in the sense of convergence of mixed moments to those of free random variables. The constructed sequence of noncommutative probability spaces is called the hierarchy of freeness with the corresponding states called m-free products of states. The first-order approximation corresponding to m=1 gives the Boolean product of states. Thus the product states for the main types of noncommutative independence can be obtained from tensor products of states. Therefore, this approach can be viewed as a unification of independence. We also show how to associate a cocommutative *-bialgebra with the m-free product of states. This approach is developed in the paper for the more general case of the conditionally free product of states and can be easily extended to families of sequences of states.


The main themes of this paper consist in showing that the principal types of noncommutative independence can be obtained from tensor independence and are also associated with certain *-bialgebras. Let \((_{*\ell\in I}{\mathcal A}, _{*\ell\in I}(\phi_\ell, \psi_\ell))\) be the conditionally free product of unital free *-algebras \({\mathcal A}_\ell\), where \(\phi_\ell\), \(\psi_\ell\) are states on \({\mathcal A}_\ell\), \(\ell\in I\). Then a sequence of noncommutative probability spaces \((\widetilde{\mathcal A}^{(m)}, \widetilde\Phi^{(m)})\), \(m\in\mathbb{N}\), is constructed with appropriate unital *-homomorphisms \(j^{(m)}: _{*\ell\in I}{\mathcal A}_\ell\to \widetilde{\mathcal A}^{(m)}\), i.e., \[ \widetilde{\mathcal A}^{(m)}= \bigotimes_{\ell\in I} \widetilde{\mathcal A}^{\otimes m}_\ell, \] \(\widetilde{\mathcal A}= {\mathcal A}*\mathbb{C}[t]\) (the free product of \({\mathcal A}\) and the algebra \(\mathbb{C}[t]\) of polynomials in one Hermitian variable \(t\)), and the tensor product state is given by \[ \widetilde\Phi^{(m)}= \bigotimes_{\ell\in I} \widetilde\phi_\ell\otimes \widetilde\psi^{\otimes(m- 1)}_\ell,\quad m\in\mathbb{N}, \] where the states \(\widetilde\phi_\ell\) (resp. \(\widetilde\psi_\ell\)) are Boolean extensions of \(\phi_\ell\) (resp. \(\psi_\ell\)), \(\ell\in I\). In fact, the main theorem says that the \(m\)-free product state \(\widetilde\Phi^{(m)}\circ j^{(m)}\) agrees with the conditionally free product state on word products \(w_1\cdots w_n\), (\(w_i\in {\mathcal A}_{k_i}\)) of not more than \(2m\) words. As a corollary, the author obtains pointwise convergence of \(\widetilde\Phi^{(m)}\circ j^{(m)}\) to the conditionally free product of states \(_{*\ell\in\{1,2\}}(\phi_i, \psi_i)\) on \(_{*i\in \{1,2\}}{\mathcal A}_i\). Since all finite joint correlations for known examples of independence can be obtained from tensor products of appropriate *-algebras, this approach can be viewed as a unification of independence. Finally, the author shows how to make the \(m\)-fold free product \(\widetilde{\mathcal A}^{*(m)}\) into a cocommutative *-bialgebra associated with \(m\)-freeness.




Cited in
(27)








This page was built for publication: Unification of Independence in Quantum Probability

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4236737)