Conditioned square functions for noncommutative martingales

Noncommutative function spaces (46L52) Noncommutative probability and statistics (46L53) Derivations, dissipations and positive semigroups in (C^*)-algebras (46L57) Martingales with discrete parameter (60G42)

Abstract: We prove a weak-type (1, 1) inequality involving conditioned versions of square functions for martingales in noncommutative

L^{p}

-spaces associated with finite von Neumann algebras. As application, we determine the optimal orders for the best constants in the noncommutative Burkholder/Rosenthal inequalities from [Ann. Probab. 31 (2003) 948--995]. We also discuss BMO-norms of sums of noncommuting order-independent operators.

Review text

This paper is the conditioned version of the author's previous work on the weak type \((1,1)\) Burkholder--Gundy inequality for noncommutative martingales [Proc. Lond. Math. Soc., III. Ser. 91, No.~2, 509--544 (2005; Zbl 1091.46040)]. The noncommutative Burkholder inequalities on the conditioned square functions were proved by \textit{M. Junge} and the reviewer for any \(1<p<\infty\) [Ann. Probab. 31, No.~2, 948--995 (2003; Zbl 1041.46050)]. These inequalities fail, of course, for \(p=1\). In the paper under review, the author obtains a weak type \((1,1)\) substitute. As a consequence, he improves certain estimates of the constants in the noncommutative Burkholder inequalities for \(1<p<\infty\) quoted above (some of them being optimal). As in the author's previous work, the arguments are based on Cuculescu's lemma for noncommutative martingales. They are, however, much more technical and subtler. The paper also includes an interesting Rosenthal type inequality in noncommutative BMO for independent random variables.

Recommendations

Cites work

Cited in

(43)

This page was built for publication: Conditioned square functions for noncommutative martingales

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2370096)