Min-entropy explained
The min-entropy, in information theory, is the smallest of the Rényi family of entropies, corresponding to the most conservative way of measuring the unpredictability of a set of outcomes, as the negative logarithm of the probability of the most likely outcome. The various Rényi entropies are all equal for a uniform distribution, but measure the unpredictability of a nonuniform distribution in different ways. The min-entropy is never greater than the ordinary or Shannon entropy (which measures the average unpredictability of the outcomes) and that in turn is never greater than the Hartley or max-entropy, defined as the logarithm of the number of outcomes with nonzero probability.
As with the classical Shannon entropy and its quantum generalization, the von Neumann entropy, one can define a conditional version of min-entropy. The conditional quantum min-entropy is a one-shot, or conservative, analog of conditional quantum entropy.
To interpret a conditional information measure, suppose Alice and Bob were to share a bipartite quantum state
. Alice has access to system
and Bob to system
. The conditional entropy measures the average uncertainty Bob has about Alice's state upon sampling from his own system. The min-entropy can be interpreted as the distance of a state from a maximally entangled state.
This concept is useful in quantum cryptography, in the context of privacy amplification (See for example [1]).
Definition for classical distributions
If
is a classical finite probability distribution, its min-entropy can be defined as
[2] One way to justify the name of the quantity is to compare it with the more standard definition of entropy, which reads
H(\boldsymbolP)=\sumipilog(1/pi)
, and can thus be written concisely as the expectation value of
over the distribution. If instead of taking the expectation value of this quantity we take its minimum value, we get precisely the above definition of
.
From an operational perspective, the min-entropy equals the negative logarithm of the probability of successfully guessing the outcome of a random draw from
.This is because it is optimal to guess the element with the largest probability and the chance of success equals the probability of that element.
Definition for quantum states
A natural way to generalize "min-entropy" from classical to quantum states is to leverage the simple observation that quantum states define classical probability distributions when measured in some basis. There is however the added difficulty that a single quantum state can result in infinitely many possible probability distributions, depending on how it is measured. A natural path is then, given a quantum state
, to still define
as
, but this time defining
as the maximum possible probability that can be obtained measuring
, maximizing over all possible projective measurements.Using this, one gets the operational definition that the min-entropy of
equals the negative logarithm of the probability of successfully guessing the outcome of any measurement of
.
Formally, this leads to the definition where we are maximizing over the set of all projective measurements
,
represent the measurement outcomes in the
POVM formalism, and
\operatorname{tr}(\Pii\rho)
is therefore the probability of observing the
-th outcome when the measurement is
.
A more concise method to write the double maximization is to observe that any element of any POVM is a Hermitian operator such that
, and thus we can equivalently directly maximize over these to get
In fact, this maximization can be performed explicitly and the maximum is obtained when
is the projection onto (any of) the largest eigenvalue(s) of
. We thus get yet another expression for the min-entropy as:
remembering that the operator norm of a Hermitian positive semidefinite operator equals its largest eigenvalue.
Conditional entropies
Let
be a bipartite density operator on the space
. The min-entropy of
conditioned on
is defined to be
Hmin(A|B)\rho\equiv
Dmax(\rhoAB\|IA ⊗ \sigmaB)
where the infimum ranges over all density operators
on the space
. The measure
is the maximum relative entropy defined as
Dmax(\rho\|\sigma)=infλ\{λ:\rho\leq2λ\sigma\}
The smooth min-entropy is defined in terms of the min-entropy.
(A|B)\rho=\sup\rho'Hmin(A|B)\rho'
where the sup and inf range over density operators
which are
-close to
. This measure of
-close is defined in terms of the purified distance
P(\rho,\sigma)=\sqrt{1-F(\rho,\sigma)2}
where
is the
fidelity measure.
These quantities can be seen as generalizations of the von Neumann entropy. Indeed, the von Neumann entropy can be expressed as
S(A|B)\rho=\lim\epsilon → \limn → infty
(An|B
~.
This is called the fully quantum asymptotic equipartition theorem.
[3] The smoothed entropies share many interesting properties with the von Neumann entropy. For example, the smooth min-entropy satisfy a data-processing inequality:
[4]
(A|B)\rho\geq
(A|BC)\rho~.
Operational interpretation of smoothed min-entropy
Henceforth, we shall drop the subscript
from the min-entropy when it is obvious from the context on what state it is evaluated.
Min-entropy as uncertainty about classical information
Suppose an agent had access to a quantum system
whose state
depends on some classical variable
. Furthermore, suppose that each of its elements
is distributed according to some distribution
. This can be described by the following state over the system
.
\rhoXB=\sumxPX(x)|x\rangle\langlex| ⊗
,
where
form an orthonormal basis. We would like to know what the agent can learn about the classical variable
. Let
be the probability that the agent guesses
when using an optimal measurement strategy
pg(X|B)=\sumxPX(x)\operatorname{tr}(Ex
,
where
is the POVM that maximizes this expression. It can be shown
[5] that this optimum can be expressed in terms of the min-entropy as
If the state
is a product state i.e.
for some density operators
and
, then there is no correlation between the systems
and
. In this case, it turns out that
Since the conditional min-entropy is always smaller than the conditional Von Neumann entropy, it follows that
Min-entropy as overlap with the maximally entangled state
The maximally entangled state
on a bipartite system
is defined as
} \sum_ |x_A\rangle |x_B\rangle
where
and
form an orthonormal basis for the spaces
and
respectively.For a bipartite quantum state
, we define the maximum overlap with the maximally entangled state as
} F\left((I_A \otimes \mathcal) \rho_, |\phi^+\rangle\langle \phi^|\right)^2
where the maximum is over all CPTP operations
and
is the dimension of subsystem
. This is a measure of how correlated the state
is. It can be shown that
. If the information contained in
is classical, this reduces to the expression above for the guessing probability.
Proof of operational characterization of min-entropy
The proof is from a paper by König, Schaffner, Renner in 2008.[6] It involves the machinery of semidefinite programs.[7] Suppose we are given some bipartite density operator
. From the definition of the min-entropy, we have
Hmin(A|B)=-
infλ\{λ|\rhoAB\leq2λ(IA ⊗ \sigmaB)\}~.
This can be re-written as
-log
\operatorname{Tr}(\sigmaB)
subject to the conditions
We notice that the infimum is taken over compact sets and hence can be replaced by a minimum. This can then be expressed succinctly as a semidefinite program. Consider the primal problem
min:\operatorname{Tr}(\sigmaB)
subjectto:IA ⊗ \sigmaB\geq\rhoAB
This primal problem can also be fully specified by the matrices
where
is the adjoint of the partial trace over
. The action of
on operators on
can be written as
\operatorname{Tr}*(X)=IA ⊗ X~.
We can express the dual problem as a maximization over operators
on the space
as
max:\operatorname{Tr}(\rhoABEAB)
subjectto:\operatorname{Tr}A(EAB)=IB
Using the Choi–Jamiołkowski isomorphism, we can define the channel
such that
dAIA ⊗ l{E}\dagger(|\phi+\rangle\langle\phi+|)=EAB
where the bell state is defined over the space
. This means that we can express the objective function of the dual problem as
\langle\rhoAB,EAB\rangle=dA\langle\rhoAB,IA ⊗ l{E}\dagger(|\phi+\rangle\langle\phi+|)\rangle
=dA\langleIA ⊗ l{E}(\rhoAB),|\phi+\rangle\langle\phi+|)\rangle
as desired.
Notice that in the event that the system
is a partly classical state as above, then the quantity that we are after reduces to
maxPX(x)\langlex|
\rangle~.
We can interpret
as a guessing strategy and this then reduces to the interpretation given above where an adversary wants to find the string
given access to quantum information via system
.
See also
Notes and References
- Vazirani . Umesh . Vidick . Thomas . Fully Device-Independent Quantum Key Distribution . Physical Review Letters . 113 . 14 . 29 September 2014 . 0031-9007 . 10.1103/physrevlett.113.140501 . 140501. 1210.1810 . 25325625. 2014PhRvL.113n0501V . 119299119 .
- König. Robert. Renner. Renato. Renato Renner. Schaffner. Christian. 2009. The Operational Meaning of Min- and Max-Entropy. IEEE Transactions on Information Theory. Institute of Electrical and Electronics Engineers (IEEE). 55. 9. 4337–4347. 0807.1338. 10.1109/tit.2009.2025545. 0018-9448. 17160454.
- Tomamichel . Marco . Colbeck . Roger . Renner . Renato . A Fully Quantum Asymptotic Equipartition Property . IEEE Transactions on Information Theory . Institute of Electrical and Electronics Engineers (IEEE) . 55 . 12 . 2009 . 0018-9448 . 10.1109/tit.2009.2032797 . 5840–5847. 0811.1221 . 12062282 .
- Renato Renner, "Security of Quantum Key Distribution", Ph.D. Thesis, Diss. ETH No. 16242
- König. Robert. Renner. Renato. Renato Renner. Schaffner. Christian. 2009. The Operational Meaning of Min- and Max-Entropy. IEEE Transactions on Information Theory. Institute of Electrical and Electronics Engineers (IEEE). 55. 9. 4337–4347. 0807.1338. 10.1109/tit.2009.2025545. 0018-9448. 17160454.
- König. Robert. Renner. Renato. Renato Renner. Schaffner. Christian. 2009. The Operational Meaning of Min- and Max-Entropy. IEEE Transactions on Information Theory. Institute of Electrical and Electronics Engineers (IEEE). 55. 9. 4337–4347. 0807.1338. 10.1109/tit.2009.2025545. 0018-9448. 17160454.
- John Watrous, Theory of quantum information, Fall 2011, course notes, https://cs.uwaterloo.ca/~watrous/CS766/LectureNotes/07.pdf