Multinomial identities arising from free probability theory. (Q1865406)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Multinomial identities arising from free probability theory.
scientific article

    Statements

    Multinomial identities arising from free probability theory. (English)
    0 references
    0 references
    26 March 2003
    0 references
    The aim of the paper is to prove a conjecture of \textit{K. Dykema} and \textit{K. Haagerup} [Am. J. Math. 126, No. 1, 121--189 (2004; Zbl 1054.47026)] saying that the moments of the triangular operator \(T,\) originally defined as a limit of distributions of normalized upper-triangular Gaussian random matrices, fulfil the relation \(\phi((T^k T^{\ast k})^n) = n^{nk}/(nk + 1)!\) for \(k,n \in \mathbb{N}.\) This result implies various other identities for multinomial coefficients. In the context of operator-valued probability spaces resp. of free probability theory, the operator \(T\) is embedded into a suitable algebra: let \(\mathcal{B} = \mathbb{C} [x],\) let \(\mathcal{A}\) denote the \(\ast\)-algebra generated by \(\mathcal{B}\) and \(\{T, T^\ast \}\) where \(T\) is defined by its variances, let \(\mathbb{E}\) denote the conditional expectation w.r.t. \(\mathcal{B}\) and define the state \(\phi\) by \(\phi(b) = \int^1_0 b(t)\, dt\), \(b \in \mathcal{B}\), \(\phi(a) = \phi (\mathbb{E} (a))\), \(a \in \mathcal{A}.\) The above mentioned conjecture follows from a more general result (Theorem 2) on multinomial coefficients \(\phi(T^{s_1}, \ldots, T^{s_n})\), \(s_i \in \{1, \ast\}\), \(n\in \mathbb{N}.\)
    0 references
    0 references
    free probability
    0 references
    Gaussian random matrices
    0 references
    multinomial coefficients
    0 references
    0 references