No label defined (Q5701322)

This is a survey article on lectures given by the author at the Grenoble summer school in 1998. The article consists of three sections. The first introductory section is devoted to a quick review of basic definitions and fundamental facts on non-commutative probability theory, including non-commutative probability spaces \(( {\mathcal A}, \varphi)\), freeness, the full Fock space \({\mathcal F}({\mathcal H})\) over a Hilbert space \({\mathcal H}\) with the vacuum \(\Omega\), the left annihilation operator \(\ell(f)\) and the left creation operator \(\ell^*(f)\) for \(f \in {\mathcal H}\), Gaussian random matrices, etc. A good reference for this part can be found, for instance, in [\textit{D. V. Voiculescu, K. J. Dykema} and \textit{A. Nica}, ``Free random variables. A noncommutative probability approach to free products with applications to random matrices, operator algebras and harmonic analysis on free groups'' (CRM Monograph Series 1, American Mathematical Society) (1992; Zbl 0795.46049)]. Section 1 treats combinatorial aspects of freeness, where the emphasis is put in particular on the concept of cumulants. Since ``freeness'' of random variables in non-commutative probability theory is defined in terms of mixed moments (having a special situation that alternating and centered moments have to vanish), it is not easy to handle this in concrete calculations. For this reason, in this article the author presents another approach to freeness more combinatorial in nature. The main emphasis is put on the so-called ``free cumulants''. Like a combinatorial description of the classical cumulants, free cumulants can also be described combinatorially, and the only difference to the classical case is that one has to replace all partitions by ``non-crossing partitions''. This combinatorial description of freeness is due to a series of works done by the author himself [Probab.\ Theory Relat.\ Fields 84, No.~2, 141--159 (1990; Zbl 0671.60109)], [Math.\ Ann.\ 298, No.~4, 611--628 (1994; Zbl 0791.06010)] and [Mem.\ Am.\ Math.\ Soc.\ 132, No.~627 (1998; Zbl 0935.46056)]; see also \textit{A.~Nica} [J.\ Funct.\ Anal.\ 135, No.~2, 271--296 (1996; Zbl 0837.60008)]. Via Voiculescu's formula related to the \(R\)-transform \(R(z)\) and Cauchy transform \(G(z)\), it is checked that the cumulants \(k_n^a\) of a random variable \(a\) are indeed the coefficients of the \(R\)-transform of \(a\) (as introduced by Voiculescu), where the distribution of \(a\) is modelled by the canonical variable (special formal power series in an isometry), see \textit{D.~V.\ Voiculescu} et al.\ [ibid.\ (1992; Zbl 0795.46049)]. Many examples are given to illustrate each abstract notion and theorem presented in this section. Section 2 is highlighted by a systematic treatment of free stochastic calculus. In this section, the author develops the analogue of a stochastic calculus for free Brownian motion. Free Brownian motion is characterized similarly to classical Brownian motion, and one only has to replace independent increments by ``free increments'' and the normal distribution by ``the semi-circle law''. In the same way as classical Brownian motion can be written as \(a_t + a_t^*\) with annihilation (respectively, creation) operators \(a_t\) (respectively \(a_t^*\)) on the Bosonic Fock space, the free Brownian motion has a canonical realization as \(\ell_t + \ell_t^*\) for left annihilation (respectively, creation) operators \(\ell_t\) (respectively \(\ell_t^*\)) on the full Fock space. Thus, instead of developing a stochastic calculus for free Brownian motion \(S_t = \ell_t + \ell_t^*\), one could also split \(S_t\) into its two summands and develop a free stochastic calculus for \(\ell_t\) and \(\ell_t^*\) in analogy to the Hudson--Parthasarathy calculus for \(a_t\) and \(a_t^*\). This was done by \textit{B.~Kümmerer} and \textit{R.~Speicher} [J.\ Funct.\ Anal.\ 103, No.~2, 372--408 (1992; Zbl 0787.46052)], while the free stochastic calculus with respect to \(S_t\) is due to \textit{P.~Biane} and \textit{R.~Speicher} [Probab.\ Theory Relat.\ Fields 112, No.~3, 373--409 (1998; Zbl 0919.60056)]. This note is based on the latter theory because it has some advantages. There are two main points to be emphasized in this presentation. The first one is that the free calculus has the very strong feature that one has estimates with respect to the uniform operator norm, i.e., the free stochastic integrals can be defined in \(L^p\) with \(p= \infty\). The second point consists in the following: free stochastic calculus is similarly characterized by its Itô formula, but is due to non-commutativity. The following items are treated in this section, such as free Brownian motion \(S_t\), stochastic integral, Itô isometry of the integral in \(L^2\)-norm, Burkholder--Gundy inequality, Itô formula in product form for free Brownian motion, Itô formula in functional form, etc. For the entire collection see [Zbl 1073.81003].