Random regularization of Brown spectral measure (Q1849073): Difference between revisions

The author proves that a convolution with a standard Gaussian random matrix regularizes the behavior of Fuglede-Kadison determinant [cf. \textit{B. Fuglede} and \textit{R. V. Kadison}, Ann. Math. (2), 55, 520-530 (1952; Zbl 0046.33604)] and Brown spectral distribution measure [cf. \textit{U. Haagerup} and \textit{F. Larsen}, J. Funct. Anal. 176, 331-367 (2000; Zbl 0984.46042)]. In this way, the author shows that it is possible to add a small random correction to a sequence of random matrices \((A^{(N)})\) which converges in *-moments almost surely to some element \(x\) in a non-commutative probability space in such a way that the new corrected sequence still converge to \(x\) and also that the empirical eigenvalue distributions of the new sequence converge to the Brown measure of \(x\) almost surely. A nice form of the random correction as a Gaussian random matrix (treated in this paper) helps the author greatly with his getting better estimates of the correction even in the operator norm. And this Gaussianity of the correction makes it easier to find the limit empirical eigenvalue distribution of a wide class of random matrices.