Generalized Stein lemma for fully quantum channel resources
- Fields
- Topics
Problem
Let \(\mathcal N:\mathcal L(A)\to\mathcal L(B)\) be a finite-dimensional quantum channel. For each \(n\), let \(\mathfrak F_n\) be a nonempty compact, convex, permutation-invariant set of channels from \(A^{\otimes n}\) to \(B^{\otimes n}\), closed under tensor products and containing \(\mathcal R_\omega^{\otimes n}\) for one full-rank state \(\omega\), where \(\mathcal R_\omega(X):=\operatorname{Tr}(X)\omega\). Define
where the supremum in Eq. (1) is over density operators and \(D\) is the quantum relative entropy. The distance to the free set is
Equation (2) is the \(n\)-use relative-entropy distance to the free channel set. For \(\varepsilon\in(0,1)\), define the optimal worst-case type-II error of a parallel quantum-input/quantum-output test by
Under what additional structural assumptions on \((\mathfrak F_n)_{n\geq1}\), if any, do both limits exist and obey the fully quantum generalized Stein identity
Source
The fully quantum formulation is implicit in the two generalized Stein theorems for classical-input channels, both of which isolate their classical-input structure from the unresolved quantum-input setting [HY25b], [BDK25].
Progress
Reports do not certify correctness or automatically change the problem's status. Progress policy.
For an i.i.d. resource state tested against admissible composite sets of free states, the generalized quantum Stein lemma identifies the fixed-error exponent with the regularized relative entropy of resource [HY25a].
Two independent works prove the channel analogue for classical–quantum channels. In that setting the channel relative entropy reduces to
\begin{equation} D_{\rm CQ}(\Phi\|\Psi)=\max_x D(\rho_x\|\sigma_x), \qquad \Phi:x\mapsto\rho_x,\quad\Psi:x\mapsto\sigma_x, \tag{5} \end{equation}and its pointwise structure supplies the additivity and minimax steps needed for a fixed-error theorem. These steps apply to Eq. (5) but are unavailable in this form for QQ channels [HY25b], [BDK25].
The CQ proof explicitly states that analogous properties for fully quantum channels remain unclear. Entangled quantum inputs introduce a reference system and a nontrivial order between input optimization and the worst-case free-channel optimization in Eq. (3) [HY25b].
Comment
The state theorem and the CQ-channel theorems do not imply Eq. (4) for genuinely quantum inputs. An adaptive quantum-comb version would be a further problem and is not included in the present statement.