Characterisations of Matrix and Operator-Valued $\Phi$-Entropies, and Operator Efron-Stein Inequalities

TL;DR

This paper develops new characterisations of matrix and operator-valued $\Phi$-entropies, proves their subadditivity, and applies these results to quantum information theory and operator variance bounds.

Contribution

It introduces operator-valued $\Phi$-entropies, establishes their subadditivity, and connects these properties to quantum channel monotonicity and variance inequalities.

Findings

Matrix $\Phi$-entropies have equivalent characterisations.

Operator-valued $\Phi$-entropies are subadditive under L"owner ordering.

Derived an operator Efron-Stein inequality for random matrices.

Abstract

We derive new characterisations of the matrix $\mathrm{\Phi}$-entropy functionals introduced in [Electron.~J.~Probab., 19(20): 1--30, 2014]. Notably, all known equivalent characterisations of the classical $\Phi$-entropies have their matrix correspondences. Next, we propose an operator-valued generalisation of the matrix $\Phi$-entropy functionals, and prove their subadditivity under L\"owner partial ordering. Our results demonstrate that the subadditivity of operator-valued $\Phi$-entropies is equivalent to the convexity of various related functions. This result can be used to demonstrate an interesting result in quantum information theory: the matrix $\Phi$-entropy of a quantum ensemble is monotone under unital quantum channels. Finally, we derive the operator Efron-Stein inequality to bound the operator-valued variance of a random matrix.