Hypercontractivity in noncommutative \(L_1\) spaces (Q1284053): Difference between revisions

In previous papers the authors have shown that given a Gibbs state \(\omega\) on an inductive limit \(C^*\)-algebra \(A\) one can construct an interpolation family of \(L_p(\omega)\) spaces, \(p\in[1,\infty]\). For any \(1\leq p\leq q\leq\infty\) one has that \(A\subset L_p\subset L_q\) and thus it is natural to consider \(A\) as a noncommutative analogue of continuous functions. Such a family provides a useful framework for construction and study of Markov semigroups, i.e. positivity and unit preserving semigroups \(P_t\), \(t\geq 0\), which are contractive in \(L_p\) spaces. It is well known that in the commutative case one can prove ergodicity for a large class of dynamics by using a technique based on the hypercontractivity of a Markov semigroup, that is \(\| P_tf\|_{L_q}\leq\| f\|_{L_p}\) for \(q\equiv q(t)= 1+(p- 1)e^{(2/c)t}\) defined with some constant \(c\in(0,\infty)\). In the noncommutative setting the hypercontractivity property has been studied only on some classes of \(L_p\) spaces associated to a trace. The present paper is the first attempt to generalize this theory to a more general setting involving the interpolating family \(L_p(\omega)\) associated to a nontrivial state \(\omega\).