A transference method in quantum probability (Q986083): Difference between revisions

Within quantum probability, it is the theory of free probability which has been developed to the greatest depth. In particular, as the authors put it: ``In terms of norm estimates for sums of independent variables, free probability often plays the role of the best of all worlds.'' There are many notions of independence in quantum probability which generalize freeness. Considered here are symmetry, subsymmetry and top-subsymmetry, all based in various ways on a general notion of independence over a von Neumann subalgebra defined by factorization of the conditional expectation. The transference method proposed and worked out in this paper is designed to allow the comparison of the \(p\)-norm of the sum of independent random variables with the \(p\)-norm of sums of free variables. Results and techniques are based on recent work in noncommutative \(L_p\)-spaces and operator spaces and developed further. For example, one of the initial ideas consists in constructing an embedding of \(L_p\) into \(L_1\) which preserves independence, and then to apply already existing results for \(p=1\) from [\textit{M. Junge}, ``Operator spaces and Araki-Woods factors: a quantum probabilistic approach'', IMRP, Int. Math. Res. Pap. 2006, No. 17, Article ID 76978 (2006; Zbl 1218.46039)]. Many of the results demand a high level of technical complexity. Applications include new results about embeddings of \(L_p\)-spaces and about Khintchine/Rosenthal inequalities for which notions of independence beyond freeness are enough. The authors pose as an open question whether the transference method can be extended beyond independent copies.