Rademacher averages on noncommutative symmetric spaces (Q999769)

The authors prove interesting and important results on Khintchine type inequalities in noncommutative Banach function spaces. Their results are best described by the abstract: ``Let \(E\) be a separable (or the dual of a separable) symmetric function space, let \(M\) be a semifinite von Neumann algebra and let \(E(M)\) be the associated noncommutative function space. Let \((\varepsilon_k)_{k\geq1}\) be a Rademacher sequence, on some probability space \(\Omega\). For finite sequences \((x_k)_{k\geq1}\) of \(E(M)\), we consider the Rademacher averages \(\sum_k\varepsilon_k\otimes x_k\) as elements of the noncommutative function space \(E(L^\infty(\Omega)\overline\otimes M)\) and study estimates for their norms \(\|\sum_k\varepsilon_k\otimes x_k\|_E\) calculated in that space. We establish general Khintchine type inequalities in this context. Then we show that if \(E\) is \(2\)-concave, \(\|\sum_k\varepsilon_k\otimes x_k\|_E\) is equivalent to the infimum of \(\|(\sum y_k^*y_k)^{1/2}\|+\|(\sum z_kz_k^*)^{1/2}\|\) over all \(y_k\), \(z_k\) in \(E(M)\) such that \(x_k=y_k+z_k\) for any \(k\geq1\). Dual estimates are given when \(E\) is \(2\)-convex and has a nontrivial upper Boyd index. In this case, \(\|\sum_k\varepsilon_k\otimes x_k\|_E\) is equivalent to \(\|(\sum x_k^*x_k)^{1/2}\|+\|(\sum x_kx_k^*)^{1/2}\|\). We also study Rademacher averages \(\sum_{i,j}\varepsilon_i\otimes\varepsilon_j\otimes x_{ij}\) for doubly indexed families \((x_{ij})_{i,j}\) of \(E(M)\).'' It should be noted that the averages considered in the paper are not the `usual' Rademacher averages, given by \[ \left\|\sum_k\varepsilon_k\otimes x_k\right\|_{\text{Rad}(E)}=\left(\int_\Omega\left\|\sum_k\varepsilon_k(\omega)x_k\right\|_{E(M)}dm(\omega)\right), \] where \(m\) is the normalized Haar measure on \(\Omega=\{-1,1\}^\infty\). In general, as discussed in the paper, the averages \(\|\cdot\|_{\text{Rad}(E)}\) and \(\|\cdot\|_E\) are not equivalent.