Let
Φ be a super-operator, i.e., a linear mapping of the form
Φ:L(F)→L(G) for finite dimensional Hilbert spaces
F and
G. This paper considers basic properties of the super-operator norms defined by
∥Φ∥q→p=sup{∥Φ(X)∥p/∥X∥q:X=0}, induced by Schatten norms for
1≤p,q≤∞. These super-operator norms arise in various contexts in the study of quantum information. In this paper it is proved that if
Φ is completely positive, the value of the supremum in the definition of
∥Φ∥q→p is achieved by a positive semidefinite operator
X, answering a question recently posed by King and Ruskai~\cite{KingR04}. However, for any choice of
p∈[1,∞], there exists a super-operator
Φ that is the {\em difference} of two completely positive, trace-preserving super-operators such that all Hermitian
X fail to achieve the supremum in the definition of
∥Φ∥1→p. Also considered are the properties of the above norms for super-operators tensored with the identity super-operator. In particular, it is proved that for all
p≥2,
q≤2, and arbitrary
Φ, the norm
∥Φ∥q→p is stable under tensoring
Φ with the identity super-operator, meaning that
∥Φ∥q→p=∥Φ⊗I∥q→p. For
1≤p<2, the norm
∥Φ∥1→p may fail to be stable with respect to tensoring
Φ with the identity super-operator as just described, but
∥Φ⊗I∥1→p is stable in this sense for
I the identity super-operator on
L(H) for $\op{dim}(\mathcal{H}) = \op{dim}(\mathcal{F})$. This generalizes and simplifies a proof due to Kitaev \cite{Kitaev97} that established this fact for the case
p=1.
For positive semidefinite matrices
A,B and
1/2≤α≤1, we establish
Tr[ADlog(Aα+Bα)(Bα)]≤ss(1−s)1−sTr[AαB1−α], where
s=(1−α)/α, with the optimal coefficient. The proof uses a superoperator representation of the directional derivative of the logarithm and an Araki-type estimate for the derivative of a fractional power. Both this inequality and its symmetric-quotient counterpart extend to positive trace-class
A and positive, possibly unbounded, self-adjoint
B on a separable Hilbert space.
The main result of the paper states that if
0=v∈Cm⊗Cn,
d=min{m,n}, and
p=SR(v), then for every
1≤q≤d, \begin{equation*} \min\left\{OSR(\mathbf U): \begin{array}{l} \mathbf U\in\mathbb M_m\otimes\mathbb M_n\text{ is unitary}, SR(\mathbf Uv)=q \end{array}\right\} = \left\lceil\max\left\{\frac pq,\frac qp\right\}\right\rceil . \end{equation*} Here
OSR(U) denotes the operator Schmidt rank of
U. Furthermore, if
U∈Mm⊗Mn is unitary and
OSR(U)≤k, then, for every unit vector
x∈Cm⊗Cn and every
1≤r≤d, \begin{equation*} F_{\min\{kr,d\}}(\mathbf Ux)\ge F_r(x) \quad\text{and}\quad F_{\min\{kr,d\}}(x)\ge F_r(\mathbf Ux), \end{equation*} where
Fr(x) denotes the sum of the
r largest squared Schmidt coefficients of
x. This result is stronger than rank reachability as it controls the entire cumulative Schmidt profile. We also establish a majorization relation and use it to obtain sharp bounds on Schatten-norm distortion and on the change of Rényi entropies under unitaries with bounded operator Schmidt rank.
We prove that there are tripartite quantum states (constructed from random unitaries) that can lead to arbitrarily large violations of Bell inequalities for dichotomic observables. As a consequence these states can withstand an arbitrary amount of white noise before they admit a description within a local hidden variable model. This is in sharp contrast with the bipartite case, where all violations are bounded by Grothendieck’s constant. We will discuss the possibility of determining the Hilbert space dimension from the obtained violation and comment on implications for communication complexity theory. Moreover, we show that the violation obtained from generalized Greenberger-Horne-Zeilinger (GHZ) states is always bounded so that, in contrast to many other contexts, GHZ states do not lead to extremal quantum correlations in this case. In order to derive all these physical consequences, we will have to obtain new mathematical results in the theories of operator spaces and tensor norms. In particular, we will prove the existence of bounded but not completely bounded trilinear forms from commutative C*-algebras. Finally, we will relate the existence of diagonal states leading to unbounded violations with a long-standing open problem in the context of Banach algebras.
Given a linear map Φ:Mn→Mm , its multiplicity maps are defined as the family of linear maps Φ⊗idk:Mn⊗Mk→Mm⊗Mk , where idk denotes the identity on Mk . Let ‖⋅‖1 denote the trace-norm on matrices, as well as the induced trace-norm on linear maps of matrices, i.e. ‖Φ‖1=max{‖Φ(X)‖1:X∈Mn,‖X‖1=1} . A fact of fundamental importance in both operator algebras and quantum information is that ‖Φ⊗idk‖1 can grow with k . In general, the rate of growth is bounded by ‖Φ⊗idk‖1≤k‖Φ‖1 , and matrix transposition is the canonical example of a map achieving this bound. We prove that, up to an equivalence, the transpose is the unique map achieving this bound. The equivalence is given in terms of complete trace-norm isometries, and the proof relies on a particular characterization of complete trace-norm isometries regarding preservation of certain multiplication relations.We use this result to characterize the set of single-shot quantum channel discrimination games satisfying a norm relation that, operationally, implies that the game can be won with certainty using entanglement, but is hard to win without entanglement. Specifically, we show that the well-known example of such a game, involving the Werner-Holevo channels, is essentially the unique game satisfying this norm relation. This constitutes a step towards a characterization of single-shot quantum channel discrimination games with maximal gap between optimal performance of entangled and unentangled strategies.