Transcript
Page 1: Introduction to Quantum Information Processingjyard/qic710/F17/Qic710Lec20-2017.pdf3 Separable states A density matrix ๐œŒ is separable if there exist probabilities ๐‘(๐‘ฅ)and

1

Introduction to

Quantum Information ProcessingQIC 710 / CS 768 / PH 767 / CO 681 / AM 871

Jon Yard

QNC 3126

[email protected]

http://math.uwaterloo.ca/~jyard/qic710

Lecture 20 (2017)

Page 2: Introduction to Quantum Information Processingjyard/qic710/F17/Qic710Lec20-2017.pdf3 Separable states A density matrix ๐œŒ is separable if there exist probabilities ๐‘(๐‘ฅ)and

2

Entanglement

Page 3: Introduction to Quantum Information Processingjyard/qic710/F17/Qic710Lec20-2017.pdf3 Separable states A density matrix ๐œŒ is separable if there exist probabilities ๐‘(๐‘ฅ)and

3

Separable statesA density matrix ๐œŒ๐ด๐ต is separable if there exist

probabilities ๐‘(๐‘ฅ) and density matrices ๐œŒ๐‘ฅ๐ด, ๐œŒ๐‘ฅ

๐ต such that

๐œŒ๐ด๐ต =

๐‘ฅ

๐‘ ๐‘ฅ ๐œŒ๐‘ฅ๐ด โŠ—๐œŒ๐‘ฅ

๐ต .

If ๐œŒ๐ด๐ต is not separable, then it is called entangled.

Note: if ๐œŒ๐ด๐ต is separable, exists a decomposition with

๐œŒ๐‘ฅ๐ด = ๐œ“๐‘ฅ โŸจ๐œ“๐‘ฅศ

๐ด, ๐œŒ๐‘ฅ๐ต = ๐œ“๐‘ฅ โŸจ๐œ“๐‘ฅศ

๐ต.

Operational meaning: separable states can be prepared

starting with only classical correlations.

Page 4: Introduction to Quantum Information Processingjyard/qic710/F17/Qic710Lec20-2017.pdf3 Separable states A density matrix ๐œŒ is separable if there exist probabilities ๐‘(๐‘ฅ)and

4

Separable?

Theorem [Horodeckis โ€˜96]: ๐œŒ๐ด๐ต is entangled iff there

exists a positive (but not completely positive) linear map

๐’œ on โ„‚๐‘‘ร—๐‘‘ such that (๐’œ โŠ— ๐‘–๐‘‘)(๐œŒ๐ด๐ต) is not positive

semidefinite.

We have already seen examples of positive-but-not-

completely positive maps, such asโ€ฆ

Proof (Easy direction โ€“ only if): Let ๐’œ be any positive map. If

๐œŒ๐ด๐ต =

๐‘ฅ

๐‘ ๐‘ฅ ๐œŒ๐‘ฅ๐ด โŠ—๐œŒ๐‘ฅ

๐ต

is a separable density matrix, then

๐‘ฅ

๐‘ ๐‘ฅ ๐’œ(๐œŒ๐‘ฅ๐ด) โŠ— ๐œŒ๐‘ฅ

๐ต

is still positive semidefinite. Interpretation: every entangled

state is broken by some non-physical positive map.

Page 5: Introduction to Quantum Information Processingjyard/qic710/F17/Qic710Lec20-2017.pdf3 Separable states A density matrix ๐œŒ is separable if there exist probabilities ๐‘(๐‘ฅ)and

5

Separable?Example: The Werner state

๐œŒ๐ด๐ต = 1 โˆ’ ๐‘๐œ™+ โŸจ๐œ™+ศ + ศ๐œ™โˆ’โŸฉโŸจ๐œ™โˆ’ศ + ๐œ“+ โŸจ๐œ“+ศ

3+ ๐‘ ๐œ“โˆ’ โŸจ๐œ“โˆ’ศ

has a Positive Partial Transpose (PPT) ๐‘‡ โŠ— ๐‘–๐‘‘ ๐œŒ๐ด๐ต โ‰ฅ 0

iff ๐‘ โ‰ค1

2, where ๐‘‡ is the transpose map ๐‘‡ ๐‘€ = ๐‘€๐‘‡.

It turns out that the PPT test is sufficient to decide

entanglement, i.e. the Werner state is entangled iff ๐‘ > 1/2.

In fact, the PPT test is sufficient to decide whether an

arbitrary 2 ร— 2 or 2 ร— 3 density matrix is entangled.

Page 6: Introduction to Quantum Information Processingjyard/qic710/F17/Qic710Lec20-2017.pdf3 Separable states A density matrix ๐œŒ is separable if there exist probabilities ๐‘(๐‘ฅ)and

6

Separable?

Fundamental problem: Given a description of ๐œŒ๐ด๐ต, (i.e.

as a ๐‘‘2 ร— ๐‘‘2 matrix), determine whether it is separable or

entangled.

Bad news: This problem is NP-hard [Gurvits โ€™02].

Good news: There exists [BCYโ€™12] an efficient

(quasipolynomial-time exp ๐œ–โˆ’2๐‘‚(log ๐‘‘ 2) algorithm for

deciding this given a promise that ๐œŒ๐ด๐ต is either separable

or a constant distance (in โ€– โ€–2-norm) from separable.

โ€–๐œŒ โˆ’ ๐œŽโ€–2 = Tr ๐œŒ โˆ’ ๐œŽ 2

Page 7: Introduction to Quantum Information Processingjyard/qic710/F17/Qic710Lec20-2017.pdf3 Separable states A density matrix ๐œŒ is separable if there exist probabilities ๐‘(๐‘ฅ)and

7

How entangled?(brief)

Page 8: Introduction to Quantum Information Processingjyard/qic710/F17/Qic710Lec20-2017.pdf3 Separable states A density matrix ๐œŒ is separable if there exist probabilities ๐‘(๐‘ฅ)and

8

Entanglement measures

Some nice properties for such a measure to satisfy:

1) Invariant under local unitaries

2) Non-increasing under Local Operations and Classical

Communication (LOCC)

3) Monogamous

4) Additive

5) Faithful

An entanglement measure is a function ๐ธ ๐œŒ๐ด๐ต on bipartite

density matrices ๐œŒ๐ด๐ต that quantifies, in one way or another,

the amount of bipartite entanglement in ๐œŒ๐ด๐ต.

Last time, we saw two examples for pure states:

โ€ข Schmidt rank

โ€ข Entanglement entropy

Page 9: Introduction to Quantum Information Processingjyard/qic710/F17/Qic710Lec20-2017.pdf3 Separable states A density matrix ๐œŒ is separable if there exist probabilities ๐‘(๐‘ฅ)and

9

Monogamy of entanglement

Many nice entanglement measures are monogamous:

The more ๐ด is entangled with ๐ต, the less it can be

entangled with ๐ถ.

๐ธ(๐œŒ๐ด๐ต1) + ๐ธ(๐œŒ๐ด๐ต2) โ‰ค ๐ธ(๐œŒ๐ด๐ต1๐ต2).Implies that quantum correlations cannot be shared.

Application of this idea: Quantum Key Distribution.

Extreme example: ๐œŒ๐ด๐ต1๐ต2 = ๐œ™ โŸจ๐œ™ศ๐ด๐ต1 โŠ—๐œŒ๐ต2, where ๐œ™ = 00 + 11 is a Bell state

1 + 0 โ‰ค 1

Page 10: Introduction to Quantum Information Processingjyard/qic710/F17/Qic710Lec20-2017.pdf3 Separable states A density matrix ๐œŒ is separable if there exist probabilities ๐‘(๐‘ฅ)and

10

Entanglement of formation

Entanglement of formation: How much entanglement

does it take, on average, to create a single copy of ๐œŒ๐ด๐ต?

๐ธ๐น(๐œŒ๐ด๐ต) = min

๐‘ ๐‘ฅ , ๐œ“๐‘ฅ๐ด๐ต

๐‘ฅ

๐‘ ๐‘ฅ ๐‘† ๐œ“๐‘ฅ๐ด :

๐‘ฅ

๐‘ ๐‘ฅ ๐œ“๐‘ฅ ๐œ“๐‘ฅ๐ด๐ต = ๐œŒ๐ด๐ต

Faithful, not monogamous, not additiveโ€ฆ

๐ธ๐ถ ๐œŒ๐ด๐ต = lim๐‘›โ†’โˆž

1

๐‘›๐ธ๐น ๐œŒ๐ด๐ต

โŠ—๐‘› โ‰ค ๐ธ๐น(๐œŒ๐ด๐ต)

Entanglement cost: how much entanglement does it

take, per copy, to create many copies of ๐œŒ๐ด๐ต?

How much entanglement does it take to make ๐œŒ๐ด๐ต using

LOCC?

Shor โ€™01, Hastings โ€˜08: Can have ๐ธ๐ถ < ๐ธ๐น (explicit example?).

Faithful, not monogamous. Additive?

Page 11: Introduction to Quantum Information Processingjyard/qic710/F17/Qic710Lec20-2017.pdf3 Separable states A density matrix ๐œŒ is separable if there exist probabilities ๐‘(๐‘ฅ)and

11

Distillable entanglementHow much entanglement can be extracted from ๐œŒ๐ด๐ต, in

the limit of many copies?does it take, on average, to

create a single copy of ๐œŒ๐ด๐ต?

๐ธ๐ท(๐œŒ๐ด๐ต) = the largest rate ๐‘… such that, by local operations

and classical communication, Alice and Bob can produce

๐‘›๐‘… Bell states (ebits)

0โŸฉศ0 + 1 ศ1โŸฉ ๐‘›๐‘… =

๐‘ฅโˆˆ 0,1 ๐‘›๐‘…

๐‘ฅ ศ๐‘ฅโŸฉ

from ๐œŒ๐ด๐ตโŠ—๐‘›

, with vanishing errors in the limit as ๐‘› โ†’ โˆž.

Page 12: Introduction to Quantum Information Processingjyard/qic710/F17/Qic710Lec20-2017.pdf3 Separable states A density matrix ๐œŒ is separable if there exist probabilities ๐‘(๐‘ฅ)and

12

Bound entanglementThere exist โ€œbound entangled statesโ€ with ๐ธ๐ท < ๐ธ๐น[Horodeckis โ€™97]

Analogous to bound energy in thermodynamics.

Has ๐ธ๐ท = 0 since it is PPT. But it is entangled.

So ๐ธ๐ท not faithful.

Big open question: do there exist NPT bound entangled states?

Would imply ๐ธ๐ท not additive.

0 < ๐‘Ž < 1

Page 13: Introduction to Quantum Information Processingjyard/qic710/F17/Qic710Lec20-2017.pdf3 Separable states A density matrix ๐œŒ is separable if there exist probabilities ๐‘(๐‘ฅ)and

13

Squashed entanglement๐ธ๐‘ ๐‘ž ๐œŒ๐ด๐ต = inf

๐œŒ๐ด๐ต๐ถ๐ผ(๐ด; ๐ตศ๐ถ)

It is monogamous, additive and faithful!

Easy to show that ๐ธ๐‘ ๐‘ž = 0 on separable states.

We donโ€™t know how to compute itโ€ฆ

๐ผ ๐ด; ๐ต ๐ถ

Conditional mutual information

๐ผ ๐ด; ๐ต ๐ถ = ๐ป ๐ด๐ถ + ๐ป ๐ต๐ถ โˆ’ ๐ป ๐ถ โˆ’ ๐ป(๐ด๐ต๐ถ)Satisfies strong subadditivity ๐ผ ๐ด; ๐ต ๐ถ โ‰ฅ 0 (not easy proof)

Generalizes mutual information

๐ผ ๐ด; ๐ต = ๐‘† ๐ด + ๐‘† ๐ต โˆ’ ๐‘†(๐ด๐ต)

Page 14: Introduction to Quantum Information Processingjyard/qic710/F17/Qic710Lec20-2017.pdf3 Separable states A density matrix ๐œŒ is separable if there exist probabilities ๐‘(๐‘ฅ)and

State redistribution problem

๐œ“ ๐ด๐ต๐ถ๐ท

Page 15: Introduction to Quantum Information Processingjyard/qic710/F17/Qic710Lec20-2017.pdf3 Separable states A density matrix ๐œŒ is separable if there exist probabilities ๐‘(๐‘ฅ)and

State redistribution problem

๐œ“ ๐ด๐ต๐ถ๐ท

Page 16: Introduction to Quantum Information Processingjyard/qic710/F17/Qic710Lec20-2017.pdf3 Separable states A density matrix ๐œŒ is separable if there exist probabilities ๐‘(๐‘ฅ)and

Cost of state redistribution

[Devetak & Y. โ€“ PRLโ€™08]

[Y. & Devetak โ€“ IEEE TIT โ€™09]

First known operational

interpretation of

quantum conditional

mutual information

๐ผ ๐ถ; ๐ท ๐ต = ๐ป ๐ต๐ถ + ๐ป ๐ต๐ทโˆ’๐ป ๐ต๐ถ๐ท โˆ’ ๐ป(๐ต)

๐ผ ๐ถ; ๐ท ๐ต ๐ผ(๐ถ; ๐ต)

๐ป(๐ถศ๐ต)

Page 17: Introduction to Quantum Information Processingjyard/qic710/F17/Qic710Lec20-2017.pdf3 Separable states A density matrix ๐œŒ is separable if there exist probabilities ๐‘(๐‘ฅ)and

Cost of state redistribution

[Devetak & Y. โ€“ PRLโ€™08]

[Y. & Devetak โ€“ IEEE TIT โ€™09]

๐ผ(๐ถ; ๐ต)

๐ป(๐ถศ๐ต)

First known operational

interpretation of

quantum conditional

mutual information

๐ผ ๐ถ; ๐ท ๐ต = ๐ป ๐ต๐ถ + ๐ป ๐ต๐ทโˆ’๐ป ๐ต๐ถ๐ท โˆ’ ๐ป(๐ต)

๐ผ ๐ถ; ๐ท ๐ต

Page 18: Introduction to Quantum Information Processingjyard/qic710/F17/Qic710Lec20-2017.pdf3 Separable states A density matrix ๐œŒ is separable if there exist probabilities ๐‘(๐‘ฅ)and

Cost of state redistribution

[Devetak & Y. โ€“ PRLโ€™08]

[Y. & Devetak โ€“ IEEE TIT โ€™09]

๐ผ(๐ถ; ๐ต)

๐ป(๐ถศ๐ต)

First known operational

interpretation of

quantum conditional

mutual information

๐ผ ๐ถ; ๐ท ๐ต = ๐ป ๐ต๐ถ + ๐ป ๐ต๐ทโˆ’๐ป ๐ต๐ถ๐ท โˆ’ ๐ป(๐ต)

๐ผ ๐ถ; ๐ท ๐ต

Page 19: Introduction to Quantum Information Processingjyard/qic710/F17/Qic710Lec20-2017.pdf3 Separable states A density matrix ๐œŒ is separable if there exist probabilities ๐‘(๐‘ฅ)and

Optimal protocol for state

redistribution

Explains the identity ๐‘›

2๐ผ ๐ถ;๐ท ๐ต =๐‘›

2๐ผ(๐ถ; ๐ทศ๐ด)

Simple proof: decoupling via random unitaries:[Oppenheim โ€“ arXiv:0805.1065]

achieves different 1-shot quantities.

Applications:

โ€ข Proof that ๐ธ๐‘ ๐‘ž is faithful.

โ€ข Proof of existence of quasipolynomial-time

algorithm for deciding separability.

โ€ข Communication complexity

Letโ€™s see how to prove a special case:

To emphasize the role of ๐ท as a reference

system, relabel ๐ท โ†’ ๐‘…

Page 20: Introduction to Quantum Information Processingjyard/qic710/F17/Qic710Lec20-2017.pdf3 Separable states A density matrix ๐œŒ is separable if there exist probabilities ๐‘(๐‘ฅ)and

State merging

Page 21: Introduction to Quantum Information Processingjyard/qic710/F17/Qic710Lec20-2017.pdf3 Separable states A density matrix ๐œŒ is separable if there exist probabilities ๐‘(๐‘ฅ)and

State merging

Page 22: Introduction to Quantum Information Processingjyard/qic710/F17/Qic710Lec20-2017.pdf3 Separable states A density matrix ๐œŒ is separable if there exist probabilities ๐‘(๐‘ฅ)and

State merging

Page 23: Introduction to Quantum Information Processingjyard/qic710/F17/Qic710Lec20-2017.pdf3 Separable states A density matrix ๐œŒ is separable if there exist probabilities ๐‘(๐‘ฅ)and

State merging

Page 24: Introduction to Quantum Information Processingjyard/qic710/F17/Qic710Lec20-2017.pdf3 Separable states A density matrix ๐œŒ is separable if there exist probabilities ๐‘(๐‘ฅ)and

State merging


Recommended