Entropy of a quantum channel

Gilad Gour1 Mark M. Wilde2

1
Department of Mathematics and Statistics
Institute for Quantum Science and Technology
University of Calgary
Alberta, Canada T2N 1N4
2
Hearne Institute for Theoretical Phyiscs
Department of Physics and Astronomy
Center for Computation and Technology
Louisiana State University
Baton Rouge, Louisiana 70803, USA

January 14, 2019

Available as arXiv:1808.06980
QIP 2019, Boulder, Colorado, USA

Gilad Gour, Mark M. Wilde Entropy of a quantum channel January 14, 2019 1/8
Motivation

von Neumann entropy is a central concept in physics and information theory, having
a number of compelling physical interpretations.

The most fundamental notion in quantum mechanics is a quantum channel, as

quantum states, unitary evolutions, measurements, and discarding of quantum
systems can each be regarded as certain kinds of quantum channels.

An important goal is thus to define a consistent and meaningful notion of the

entropy of a quantum channel.

Gilad Gour, Mark M. Wilde Entropy of a quantum channel January 14, 2019 2/8
von Neumann entropy

Entropy of a quantum state

For a quantum state ρA of a system A, the entropy is defined as

H(A)ρ ≡ − Tr{ρA log2 ρA }.

It has operational interpretations in terms of quantum data compression and optimal

entanglement manipulation rates of pure bipartite quantum states.

Entropy in terms of quantum relative entropy

By defining the quantum relative entropy of ρA and σA as

D(ρA kσA ) = Tr{ρA [log2 ρA − log2 σA ]},

we can rewrite the formula for quantum entropy as follows:

H(A)ρ = log2 |A| − D(ρA kπA ),

where |A| is dimension of A and πA ≡ IA / |A| denotes the maximally mixed state.

Gilad Gour, Mark M. Wilde Entropy of a quantum channel January 14, 2019 3/8
Entropy of a quantum channel

The relative entropy of quantum channels NA→B and MA→B is defined as

D(N kM) ≡ sup D(NA→B (ρRA )kMA→B (ρRA )),

ρRA

where the optimization is with respect to bipartite states ρRA of a reference

system R of arbitrary size and the channel input system A.

Completely depolarizing channel generalizes the maximally mixed state:

RA→B (XA ) = Tr{XA }πB ,

where XA is an arbitrary operator for system A and πB is maximally mixed.

Entropy of a quantum channel

Let NA→B be a quantum channel. Then its entropy is defined as

H(N ) ≡ log2 |B| − D(N kR).

See also [Yuan, arXiv:1807.05958].

Gilad Gour, Mark M. Wilde Entropy of a quantum channel January 14, 2019 4/8
Properties of the entropy of a channel

Non-decrease under the action of a uniformity preserving superchannel: Let NA→B

be a quantum channel, and let Θ be a uniformity preserving superchannel (one that
preserves the completely randomizing channel). Then

H(Θ(N )) ≥ H(N ).

Additivity: Let N and M be quantum channels. Then the channel entropy is

additive in the following sense:

H(N ⊗ M) = H(N ) + H(M).

Reduction to states and normalization: Let the channel NA→B be a replacer

channel, defined such that NA→B (ρA ) = σB for all states ρA and some state σB .
Then the following equality holds

H(N ) = H(B)σ .

Gilad Gour, Mark M. Wilde Entropy of a quantum channel January 14, 2019 5/8
Rényi entropy of a quantum channel

Recall that the sandwiched Rényi relative entropy of ρ and σ is defined as

1 n α o
Dα (ρkσ) ≡ log2 Tr σ (1−α)/2α ρσ (1−α)/2α .
α−1

This leads to the sandwiched Rényi divergence of channels NA→B and MA→B :

Dα (N kM) ≡ sup Dα (NA→B (ρRA )kMA→B (ρRA )).

ρRA

Rényi entropy of a channel

Let NA→B be a quantum channel. For α ∈ [1/2, 1) ∪ (1, ∞), the Rényi entropy of the
channel N is defined as
Hα (N ) ≡ log2 |B| − Dα (N kR),
where RA→B is the completely randomizing channel. It also obeys non-decrease under a
uniformity-preserving superchannel, additivity, and reduction to states.

Gilad Gour, Mark M. Wilde Entropy of a quantum channel January 14, 2019 6/8
Asymptotic Equipartition Property

Define the purified channel divergence of two channels NA→B and MA→B as

P(N , M) ≡ sup P(NA→B (ρRA ), MA→B (ρRA )),

ρRA

p √ √
where P(ω, τ ) ≡ 1 − k ω τ k21 is the purified distance of states.

Define the smoothed min-entropy of a channel for ε ∈ (0, 1) as

ε
Hmin (N ) ≡ sup Hmin (N
e ).
P(N ,N
f)≤ε

Asymptotic equipartition property: For all ε ∈ (0, 1), the following inequality holds
1 ε
lim Hmin (N ⊗n ) ≥ H(N ).
n→∞ n
We also have that
1 ε
lim lim Hmin (N ⊗n ) ≤ H(N ).
ε→0 n→∞ n

Gilad Gour, Mark M. Wilde Entropy of a quantum channel January 14, 2019 7/8
Operational interpretation in terms of quantum channel merging

Bn Bn

An An ΦK ΦL
VN Vs. VN P
Bn Bn
En En
En

Goal: for Bob to merge his share of the channel with Eve’s.
Given channel NA→B , let V N ≡ (UA→BE
N
)⊗n , where UA→BE
N
extends NA→B .
By consuming a maximally entangled state ΦK of Schmidt rank K and applying a
one-way LOCC protocol P , Bob and Eve can distill a maximally entangled state ΦL
of Schmidt rank L and transfer Bob’s systems B n to Eve, in such a way that any
third party having access to the inputs An and the outputs B n and E n would not be
able to distinguish the difference between the ideal situation on the left and the
simulation on the right.
The optimal asymptotic rate of entanglement gain is equal to
the entropy of the channel N .

Gilad Gour, Mark M. Wilde Entropy of a quantum channel January 14, 2019 8/8