Entanglement and secret-key-agreement capacities of bipartite - - PowerPoint PPT Presentation

entanglement and secret key agreement capacities of
SMART_READER_LITE
LIVE PREVIEW

Entanglement and secret-key-agreement capacities of bipartite - - PowerPoint PPT Presentation

Entanglement and secret-key-agreement capacities of bipartite quantum interactions and read-only memory devices Siddhartha Das 1* auml 2 Mark M. Wilde 1 Stefan B 1 Louisiana State University, USA 2 Delft University of Technology, Netherlands


slide-1
SLIDE 1

Entanglement and secret-key-agreement capacities of bipartite quantum interactions and read-only memory devices

Siddhartha Das 1* Stefan B¨ auml 2 Mark M. Wilde 1

1Louisiana State University, USA 2Delft University of Technology, Netherlands & NTT Japan ∗sdas21@lsu.edu

arXiv:1712.00827

8th International Conference on Quantum Cryptography, Shanghai, China

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 1 / 16

slide-2
SLIDE 2

Bipartite quantum interactions

Bipartite unitary interactions are the most elementary many-body interactions. Due to unavoidable interaction with environment, study of bipartite noisy interactions is pertinent.

E' A' B' E A B

Figure: Systems of interest A′ and B′ interacting in presence of the bath E ′.

U is unitary transformation corresponding to underlying interaction Hamiltonian ˆ H among A′, B′, E ′. Before action of interaction Hamiltonian ˆ H: ωA′B′ ⊗ τE ′, where bath E ′ is in some fixed state and uncorrelated to A′B′. After action of ˆ H: ρABE := UA′B′E ′→ABE(ωA′B′ ⊗ τE ′).

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 2 / 16

slide-3
SLIDE 3

Bipartite quantum interactions

Bipartite unitary interactions are the most elementary many-body interactions. Due to unavoidable interaction with environment, study of bipartite noisy interactions is pertinent.

E' A' B' E A B

Figure: Systems of interest A′ and B′ interacting in presence of the bath E ′.

U is unitary transformation corresponding to underlying interaction Hamiltonian ˆ H among A′, B′, E ′. Before action of interaction Hamiltonian ˆ H: ωA′B′ ⊗ τE ′, where bath E ′ is in some fixed state and uncorrelated to A′B′. After action of ˆ H: ρABE := UA′B′E ′→ABE(ωA′B′ ⊗ τE ′).

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 2 / 16

slide-4
SLIDE 4

Bipartite quantum interactions

Bipartite unitary interactions are the most elementary many-body interactions. Due to unavoidable interaction with environment, study of bipartite noisy interactions is pertinent.

E' A' B' E A B

Figure: Systems of interest A′ and B′ interacting in presence of the bath E ′.

U is unitary transformation corresponding to underlying interaction Hamiltonian ˆ H among A′, B′, E ′. Before action of interaction Hamiltonian ˆ H: ωA′B′ ⊗ τE ′, where bath E ′ is in some fixed state and uncorrelated to A′B′. After action of ˆ H: ρABE := UA′B′E ′→ABE(ωA′B′ ⊗ τE ′).

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 2 / 16

slide-5
SLIDE 5

Bipartite quantum interactions

Bipartite unitary interactions are the most elementary many-body interactions. Due to unavoidable interaction with environment, study of bipartite noisy interactions is pertinent.

E' A' B' E A B

Figure: Systems of interest A′ and B′ interacting in presence of the bath E ′.

U is unitary transformation corresponding to underlying interaction Hamiltonian ˆ H among A′, B′, E ′. Before action of interaction Hamiltonian ˆ H: ωA′B′ ⊗ τE ′, where bath E ′ is in some fixed state and uncorrelated to A′B′. After action of ˆ H: ρABE := UA′B′E ′→ABE(ωA′B′ ⊗ τE ′).

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 2 / 16

slide-6
SLIDE 6

Bidirectional quantum channel

A bipartite quantum channel NA′B′→AB is a completely positive, trace-preserving map that transforms composite system A′B′ to AB.

A' B' A B

Figure: Two parties of interest: Alice holds A′, A and Bob holds B′, B.

When A′, A are held by Alice and B′, B are held by Bob, bipartite channel N is called bidirectional channel. It corresponds to noisy bipartite interaction, when bath is inaccessible. For all input state ωA′B′: N(ωA′B′) = ρAB, where ρAB := TrE{UA′B′E ′→ABE(ωA′B′ ⊗ τE ′)}, when initial state τE ′ of bath is fixed.

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 3 / 16

slide-7
SLIDE 7

Bidirectional quantum channel

A bipartite quantum channel NA′B′→AB is a completely positive, trace-preserving map that transforms composite system A′B′ to AB.

A' B' A B

Figure: Two parties of interest: Alice holds A′, A and Bob holds B′, B.

When A′, A are held by Alice and B′, B are held by Bob, bipartite channel N is called bidirectional channel. It corresponds to noisy bipartite interaction, when bath is inaccessible. For all input state ωA′B′: N(ωA′B′) = ρAB, where ρAB := TrE{UA′B′E ′→ABE(ωA′B′ ⊗ τE ′)}, when initial state τE ′ of bath is fixed.

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 3 / 16

slide-8
SLIDE 8

Bidirectional quantum channel

A bipartite quantum channel NA′B′→AB is a completely positive, trace-preserving map that transforms composite system A′B′ to AB.

A' B' A B

Figure: Two parties of interest: Alice holds A′, A and Bob holds B′, B.

When A′, A are held by Alice and B′, B are held by Bob, bipartite channel N is called bidirectional channel. It corresponds to noisy bipartite interaction, when bath is inaccessible. For all input state ωA′B′: N(ωA′B′) = ρAB, where ρAB := TrE{UA′B′E ′→ABE(ωA′B′ ⊗ τE ′)}, when initial state τE ′ of bath is fixed.

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 3 / 16

slide-9
SLIDE 9

Motivation

Bidirectional channels: Simple model of quantum network with 2 clients, Alice and Bob. Model for quantum gates – CNOT, SWAP, etc.– in noisy intermediate-scale quantum (NISQ) computers. Entanglement may increase, decrease or not change due to bipartite quantum interactions. Entanglement distillation: Maximally entangled states are useful resource for several information processing tasks: quantum key distribution, quantum teleportation, etc. Secret key distillation: Need for secure communication protocols between two parties over network – private reading.

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 4 / 16

slide-10
SLIDE 10

Motivation

Bidirectional channels: Simple model of quantum network with 2 clients, Alice and Bob. Model for quantum gates – CNOT, SWAP, etc.– in noisy intermediate-scale quantum (NISQ) computers. Entanglement may increase, decrease or not change due to bipartite quantum interactions. Entanglement distillation: Maximally entangled states are useful resource for several information processing tasks: quantum key distribution, quantum teleportation, etc. Secret key distillation: Need for secure communication protocols between two parties over network – private reading.

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 4 / 16

slide-11
SLIDE 11

Motivation

Bidirectional channels: Simple model of quantum network with 2 clients, Alice and Bob. Model for quantum gates – CNOT, SWAP, etc.– in noisy intermediate-scale quantum (NISQ) computers. Entanglement may increase, decrease or not change due to bipartite quantum interactions. Entanglement distillation: Maximally entangled states are useful resource for several information processing tasks: quantum key distribution, quantum teleportation, etc. Secret key distillation: Need for secure communication protocols between two parties over network – private reading.

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 4 / 16

slide-12
SLIDE 12

Motivation

Bidirectional channels: Simple model of quantum network with 2 clients, Alice and Bob. Model for quantum gates – CNOT, SWAP, etc.– in noisy intermediate-scale quantum (NISQ) computers. Entanglement may increase, decrease or not change due to bipartite quantum interactions. Entanglement distillation: Maximally entangled states are useful resource for several information processing tasks: quantum key distribution, quantum teleportation, etc. Secret key distillation: Need for secure communication protocols between two parties over network – private reading.

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 4 / 16

slide-13
SLIDE 13

Motivation

Bidirectional channels: Simple model of quantum network with 2 clients, Alice and Bob. Model for quantum gates – CNOT, SWAP, etc.– in noisy intermediate-scale quantum (NISQ) computers. Entanglement may increase, decrease or not change due to bipartite quantum interactions. Entanglement distillation: Maximally entangled states are useful resource for several information processing tasks: quantum key distribution, quantum teleportation, etc. Secret key distillation: Need for secure communication protocols between two parties over network – private reading.

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 4 / 16

slide-14
SLIDE 14

Goal

Two different information-processing tasks relevant for bipartite quantum interactions:

1

Entanglement distillation: generation of singlet state from two separated systems.

2

Secret key agreement: generation of maximal classical correlation between two separated systems, such that there’s no correlation with the bath.

New secure communication protocol between two parties, called private reading. Non-asymptotic capacity of a channel N for a task: Maximum rate at which a given task can be accomplished by allowing the use of N a finite number of times.

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 5 / 16

slide-15
SLIDE 15

Goal

Two different information-processing tasks relevant for bipartite quantum interactions:

1

Entanglement distillation: generation of singlet state from two separated systems.

2

Secret key agreement: generation of maximal classical correlation between two separated systems, such that there’s no correlation with the bath.

New secure communication protocol between two parties, called private reading. Non-asymptotic capacity of a channel N for a task: Maximum rate at which a given task can be accomplished by allowing the use of N a finite number of times.

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 5 / 16

slide-16
SLIDE 16

Goal

Two different information-processing tasks relevant for bipartite quantum interactions:

1

Entanglement distillation: generation of singlet state from two separated systems.

2

Secret key agreement: generation of maximal classical correlation between two separated systems, such that there’s no correlation with the bath.

New secure communication protocol between two parties, called private reading. Non-asymptotic capacity of a channel N for a task: Maximum rate at which a given task can be accomplished by allowing the use of N a finite number of times.

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 5 / 16

slide-17
SLIDE 17

Goal

Two different information-processing tasks relevant for bipartite quantum interactions:

1

Entanglement distillation: generation of singlet state from two separated systems.

2

Secret key agreement: generation of maximal classical correlation between two separated systems, such that there’s no correlation with the bath.

New secure communication protocol between two parties, called private reading. Non-asymptotic capacity of a channel N for a task: Maximum rate at which a given task can be accomplished by allowing the use of N a finite number of times.

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 5 / 16

slide-18
SLIDE 18

Goal

Two different information-processing tasks relevant for bipartite quantum interactions:

1

Entanglement distillation: generation of singlet state from two separated systems.

2

Secret key agreement: generation of maximal classical correlation between two separated systems, such that there’s no correlation with the bath.

New secure communication protocol between two parties, called private reading. Non-asymptotic capacity of a channel N for a task: Maximum rate at which a given task can be accomplished by allowing the use of N a finite number of times.

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 5 / 16

slide-19
SLIDE 19

Secret key generation over bidirectional channel

LOCC LOCC LOCC LOCC LOCC

LOCC-assisted bidirectional secret-key-agreement capacity [P2→2

LOCC(NA′B′→AB).]

Bidirectional max-relative entropy of entanglement: E 2→2

max (NA′B′→AB) =

sup

ψSAA′⊗ϕB′SB

Emax(SAA; BSB)N(ψ⊗ϕ), where ψSAA′, ϕB′SB are pure states, SA ≃ A′, SB ≃ B′, Emax(A : B)ρ = minσAB∈SEP Dmax(ρσ), such that Dmax(ρσ) = inf{λ : ρ ≤ 2λσ}.

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 6 / 16

slide-20
SLIDE 20

Secret key generation over bidirectional channel

LOCC LOCC LOCC LOCC LOCC

LOCC-assisted bidirectional secret-key-agreement capacity [P2→2

LOCC(NA′B′→AB).]

Bidirectional max-relative entropy of entanglement: E 2→2

max (NA′B′→AB) =

sup

ψSAA′⊗ϕB′SB

Emax(SAA; BSB)N(ψ⊗ϕ), where ψSAA′, ϕB′SB are pure states, SA ≃ A′, SB ≃ B′, Emax(A : B)ρ = minσAB∈SEP Dmax(ρσ), such that Dmax(ρσ) = inf{λ : ρ ≤ 2λσ}.

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 6 / 16

slide-21
SLIDE 21

Secret key generation over bidirectional channel

LOCC LOCC LOCC LOCC LOCC

LOCC-assisted bidirectional secret-key-agreement capacity [P2→2

LOCC(NA′B′→AB).]

Bidirectional max-relative entropy of entanglement: E 2→2

max (NA′B′→AB) =

sup

ψSAA′⊗ϕB′SB

Emax(SAA; BSB)N(ψ⊗ϕ), where ψSAA′, ϕB′SB are pure states, SA ≃ A′, SB ≃ B′, Emax(A : B)ρ = minσAB∈SEP Dmax(ρσ), such that Dmax(ρσ) = inf{λ : ρ ≤ 2λσ}.

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 6 / 16

slide-22
SLIDE 22

Secret key generation over bidirectional channel

LOCC LOCC LOCC LOCC LOCC

Theorem (LOCC-assisted secret key agreement)

1 n log2 M ≤ E 2→2

max (N) + 1

n log2

  • 1

1 − ε

  • .

P2→2

LOCC(NA′B′→AB) ≤ E 2→2 max (NA′B′→AB), and this upper bound is in fact a strong converse

bound.

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 7 / 16

slide-23
SLIDE 23

Secret key generation over bidirectional channel

LOCC LOCC LOCC LOCC LOCC

Theorem (LOCC-assisted secret key agreement)

1 n log2 M ≤ E 2→2

max (N) + 1

n log2

  • 1

1 − ε

  • .

P2→2

LOCC(NA′B′→AB) ≤ E 2→2 max (NA′B′→AB), and this upper bound is in fact a strong converse

bound.

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 7 / 16

slide-24
SLIDE 24

Secret key generation over bidirectional channel

LOCC LOCC LOCC LOCC LOCC

Theorem (LOCC-assisted secret key agreement)

1 n log2 M ≤ E 2→2

max (N) + 1

n log2

  • 1

1 − ε

  • .

P2→2

LOCC(NA′B′→AB) ≤ E 2→2 max (NA′B′→AB), and this upper bound is in fact a strong converse

bound.

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 7 / 16

slide-25
SLIDE 25

Proof outline: Secret key generation

LOCC LOCC LOCC LOCC LOCC

Private states [HHHO05,HHHO09]: State γSAKA:KBSB containing log2 K private bits. Success probability in privacy test: Tr {Πγω} ≥ 1 − ε. By [HHHO05,HHHO09], Tr {Πγσ} ≤ 1

K for σ ∈ SEP.

Main observation: E 2→2

max is not enhanced by amortization.

Emax(SAA; BSB)σ ≤ Emax(SAA′; B′SB)ρ + E 2→2

max (NA′B′→AB),

where σSAABSB = NA′B′→AB(ρSAA′B′SB).

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 8 / 16

slide-26
SLIDE 26

Proof outline: Secret key generation

LOCC LOCC LOCC LOCC LOCC

Private states [HHHO05,HHHO09]: State γSAKA:KBSB containing log2 K private bits. Success probability in privacy test: Tr {Πγω} ≥ 1 − ε. By [HHHO05,HHHO09], Tr {Πγσ} ≤ 1

K for σ ∈ SEP.

Main observation: E 2→2

max is not enhanced by amortization.

Emax(SAA; BSB)σ ≤ Emax(SAA′; B′SB)ρ + E 2→2

max (NA′B′→AB),

where σSAABSB = NA′B′→AB(ρSAA′B′SB).

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 8 / 16

slide-27
SLIDE 27

Proof outline: Secret key generation

LOCC LOCC LOCC LOCC LOCC

Private states [HHHO05,HHHO09]: State γSAKA:KBSB containing log2 K private bits. Success probability in privacy test: Tr {Πγω} ≥ 1 − ε. By [HHHO05,HHHO09], Tr {Πγσ} ≤ 1

K for σ ∈ SEP.

Main observation: E 2→2

max is not enhanced by amortization.

Emax(SAA; BSB)σ ≤ Emax(SAA′; B′SB)ρ + E 2→2

max (NA′B′→AB),

where σSAABSB = NA′B′→AB(ρSAA′B′SB).

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 8 / 16

slide-28
SLIDE 28

Proof outline: Secret key generation

LOCC LOCC LOCC LOCC LOCC

Private states [HHHO05,HHHO09]: State γSAKA:KBSB containing log2 K private bits. Success probability in privacy test: Tr {Πγω} ≥ 1 − ε. By [HHHO05,HHHO09], Tr {Πγσ} ≤ 1

K for σ ∈ SEP.

Main observation: E 2→2

max is not enhanced by amortization.

Emax(SAA; BSB)σ ≤ Emax(SAA′; B′SB)ρ + E 2→2

max (NA′B′→AB),

where σSAABSB = NA′B′→AB(ρSAA′B′SB).

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 8 / 16

slide-29
SLIDE 29

Entanglement generation over bidirectional channel

N

LA LB B’2 B2 A ’2

N

LA2 A2 LB2

PPT-P

B’1 B1 A ’1

PPT-P PPT-P

N

B’ B A

PPT-P PPT-P

n n n

A

M

B

M LA1 A1 LB1 A ’n

n n

PPT-assisted bidirectional quantum capacity

Q2→2

PPT(NA′B′→AB).

  • Bidirectional max-Rains Information

R2→2

max (NA′B′→AB) = log Γ2→2(NA′B′→AB)

, where

Γ2→2(NA′B′→AB) = minimize TrAB{VSAABSB + YSAABSB}∞ subject to VSAABSB, YSAABSB ≥ 0, TBSB{VSAABSB − YSAABSB} ≥ JN

SAABSB,

such that SA ≃ A′, SB ≃ B′.

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 9 / 16

slide-30
SLIDE 30

Entanglement generation over bidirectional channel

N

LA LB B’2 B2 A ’2

N

LA2 A2 LB2

PPT-P

B’1 B1 A ’1

PPT-P PPT-P

N

B’ B A

PPT-P PPT-P

n n n

A

M

B

M LA1 A1 LB1 A ’n

n n

PPT-assisted bidirectional quantum capacity

Q2→2

PPT(NA′B′→AB).

  • Bidirectional max-Rains Information

R2→2

max (NA′B′→AB) = log Γ2→2(NA′B′→AB)

, where

Γ2→2(NA′B′→AB) = minimize TrAB{VSAABSB + YSAABSB}∞ subject to VSAABSB, YSAABSB ≥ 0, TBSB{VSAABSB − YSAABSB} ≥ JN

SAABSB,

such that SA ≃ A′, SB ≃ B′.

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 9 / 16

slide-31
SLIDE 31

Entanglement generation over bidirectional channel

N

LA LB B’2 B2 A ’2

N

LA2 A2 LB2

PPT-P

B’1 B1 A ’1

PPT-P PPT-P

N

B’ B A

PPT-P PPT-P

n n n

A

M

B

M LA1 A1 LB1 A ’n

n n

PPT-assisted bidirectional quantum capacity

Q2→2

PPT(NA′B′→AB).

  • Bidirectional max-Rains Information

R2→2

max (NA′B′→AB) = log Γ2→2(NA′B′→AB)

, where

Γ2→2(NA′B′→AB) = minimize TrAB{VSAABSB + YSAABSB}∞ subject to VSAABSB, YSAABSB ≥ 0, TBSB{VSAABSB − YSAABSB} ≥ JN

SAABSB,

such that SA ≃ A′, SB ≃ B′.

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 9 / 16

slide-32
SLIDE 32

Entanglement generation over bidirectional channel

N

LA LB B’2 B2 A ’2

N

LA2 A2 LB2

PPT-P

B’1 B1 A ’1

PPT-P PPT-P

N

B’ B A

PPT-P PPT-P

n n n

A

M

B

M LA1 A1 LB1 A ’n

n n

PPT-assisted bidirectional quantum capacity

Q2→2

PPT(NA′B′→AB).

  • Bidirectional max-Rains Information

R2→2

max (NA′B′→AB) = log Γ2→2(NA′B′→AB)

, where

Γ2→2(NA′B′→AB) = minimize TrAB{VSAABSB + YSAABSB}∞ subject to VSAABSB, YSAABSB ≥ 0, TBSB{VSAABSB − YSAABSB} ≥ JN

SAABSB,

such that SA ≃ A′, SB ≃ B′.

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 9 / 16

slide-33
SLIDE 33

Entanglement generation over bidirectional channel

N

LA LB B’2 B2 A ’2

N

LA2 A2 LB2

PPT-P

B’1 B1 A ’1

PPT-P PPT-P

N

B’ B A

PPT-P PPT-P

n n n

A

M

B

M LA1 A1 LB1 A ’n

n n

Theorem (PPT-assisted distillable entanglement generation)

1 n log2 M ≤ R2→2

max (N) + 1

n log2

  • 1

1 − ε

  • .

Q2→2

PPT(NA′B′→AB) ≤ R2→2 max (NA′B′→AB), and this upper bound is in fact a strong converse

bound.

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 10 / 16

slide-34
SLIDE 34

Entanglement generation over bidirectional channel

N

LA LB B’2 B2 A ’2

N

LA2 A2 LB2

PPT-P

B’1 B1 A ’1

PPT-P PPT-P

N

B’ B A

PPT-P PPT-P

n n n

A

M

B

M LA1 A1 LB1 A ’n

n n

Theorem (PPT-assisted distillable entanglement generation)

1 n log2 M ≤ R2→2

max (N) + 1

n log2

  • 1

1 − ε

  • .

Q2→2

PPT(NA′B′→AB) ≤ R2→2 max (NA′B′→AB), and this upper bound is in fact a strong converse

bound.

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 10 / 16

slide-35
SLIDE 35

Entanglement generation over bidirectional channel

N

LA LB B’2 B2 A ’2

N

LA2 A2 LB2

PPT-P

B’1 B1 A ’1

PPT-P PPT-P

N

B’ B A

PPT-P PPT-P

n n n

A

M

B

M LA1 A1 LB1 A ’n

n n

Theorem (PPT-assisted distillable entanglement generation)

1 n log2 M ≤ R2→2

max (N) + 1

n log2

  • 1

1 − ε

  • .

Q2→2

PPT(NA′B′→AB) ≤ R2→2 max (NA′B′→AB), and this upper bound is in fact a strong converse

bound.

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 10 / 16

slide-36
SLIDE 36

Application: Private Reading

First recall: (Quantum) Reading [BRV00,Pir11] of memory devices. Memory device: message encoded into a sequence of channels from a memory cell SX = {N x

B′→B}x∈X .

Alice encodes m ∈ M into codewords (x1(m), ..., xn(m)), and sets the device to

  • N x1(m)

B′→B, ..., N xn(m) B′→B

  • .

Bob can enter quantum states and do channel discrimination to learn m. Natural to employ adaptive strategy [DW17].

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 11 / 16

slide-37
SLIDE 37

Application: Private Reading

First recall: (Quantum) Reading [BRV00,Pir11] of memory devices. Memory device: message encoded into a sequence of channels from a memory cell SX = {N x

B′→B}x∈X .

Alice encodes m ∈ M into codewords (x1(m), ..., xn(m)), and sets the device to

  • N x1(m)

B′→B, ..., N xn(m) B′→B

  • .

Bob can enter quantum states and do channel discrimination to learn m. Natural to employ adaptive strategy [DW17].

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 11 / 16

slide-38
SLIDE 38

Application: Private Reading

First recall: (Quantum) Reading [BRV00,Pir11] of memory devices. Memory device: message encoded into a sequence of channels from a memory cell SX = {N x

B′→B}x∈X .

Alice encodes m ∈ M into codewords (x1(m), ..., xn(m)), and sets the device to

  • N x1(m)

B′→B, ..., N xn(m) B′→B

  • .

Bob can enter quantum states and do channel discrimination to learn m. Natural to employ adaptive strategy [DW17].

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 11 / 16

slide-39
SLIDE 39

Application: Private Reading

First recall: (Quantum) Reading [BRV00,Pir11] of memory devices. Memory device: message encoded into a sequence of channels from a memory cell SX = {N x

B′→B}x∈X .

Alice encodes m ∈ M into codewords (x1(m), ..., xn(m)), and sets the device to

  • N x1(m)

B′→B, ..., N xn(m) B′→B

  • .

Bob can enter quantum states and do channel discrimination to learn m. Natural to employ adaptive strategy [DW17].

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 11 / 16

slide-40
SLIDE 40

Application: Private Reading E3 E2 ê

k

B3 B’3 B’2 B1

x1(k)

M

B’1 B2

A1

x2(k)

M

A2

x3(k)

M

L L L E1

B1 B2 B3

Private reading: Eve present when Bob performs the readout: Wiretap memory cell ¯ SX = {Mx

B′→BE}x∈X .

Special case: Isometric wiretap memory cell ¯ Siso

X .

The non-adaptive private reading capacity of a wiretap memory cell SX is given by Pread

n-a

  • MX
  • = sup

n

max

pXn,σLBB′n

1 n [I(X n; LBBn)τ − I(X n; E n)τ] , where τX nLBBnE n :=

xn pX n(xn)|xn

xn|X n ⊗ Mxn

B′n→BnE n(σLBB′n).

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 12 / 16

slide-41
SLIDE 41

Application: Private Reading E3 E2 ê

k

B3 B’3 B’2 B1

x1(k)

M

B’1 B2

A1

x2(k)

M

A2

x3(k)

M

L L L E1

B1 B2 B3

Private reading: Eve present when Bob performs the readout: Wiretap memory cell ¯ SX = {Mx

B′→BE}x∈X .

Special case: Isometric wiretap memory cell ¯ Siso

X .

The non-adaptive private reading capacity of a wiretap memory cell SX is given by Pread

n-a

  • MX
  • = sup

n

max

pXn,σLBB′n

1 n [I(X n; LBBn)τ − I(X n; E n)τ] , where τX nLBBnE n :=

xn pX n(xn)|xn

xn|X n ⊗ Mxn

B′n→BnE n(σLBB′n).

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 12 / 16

slide-42
SLIDE 42

Application: Private Reading E3 E2 ê

k

B3 B’3 B’2 B1

x1(k)

M

B’1 B2

A1

x2(k)

M

A2

x3(k)

M

L L L E1

B1 B2 B3

Private reading: Eve present when Bob performs the readout: Wiretap memory cell ¯ SX = {Mx

B′→BE}x∈X .

Special case: Isometric wiretap memory cell ¯ Siso

X .

The non-adaptive private reading capacity of a wiretap memory cell ¯ SX is given by Pread

n-a

¯

SX

  • = sup

n

max

pXn,σLBB′n

1 n [I(X n; LBBn)τ − I(X n; E n)τ] , where τX nLBBnE n :=

xn pX n(xn)|xn

xn|X n ⊗ Mxn

B′n→BnE n(σLBB′n).

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 13 / 16

slide-43
SLIDE 43

Application: Private Reading E3 E2 ê

k

B3 B’3 B’2 B1

x1(k)

M

B’1 B2

A1

x2(k)

M

A2

x3(k)

M

L L L E1

B1 B2 B3

Private reading: Eve present when Bob performs the readout: Wiretap memory cell ¯ SX = {Mx

B′→BE}x∈X .

Special case: Isometric wiretap memory cell ¯ Siso

X .

The non-adaptive private reading capacity of a wiretap memory cell ¯ SX is given by Pread

n-a

¯

SX

  • = sup

n

max

pXn,σLBB′n

1 n [I(X n; LBBn)τ − I(X n; E n)τ] , where τX nLBBnE n :=

xn pX n(xn)|xn

xn|X n ⊗ Mxn

B′n→BnE n(σLBB′n).

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 13 / 16

slide-44
SLIDE 44

Application: Private Reading E3 E2 ê

k

B3 B’3 B’2 B1

x1(k)

M

B’1 B2

A1

x2(k)

M

A2

x3(k)

M

L L L E1

B1 B2 B3

Private reading: Eve present when Bob performs the readout: Wiretap memory cell ¯ SX = {Mx

B′→BE}x∈X .

Special case: Isometric wiretap memory cell ¯ Siso

X .

The non-adaptive private reading capacity of a wiretap memory cell ¯ SX is given by Pread

n-a

¯

SX

  • = sup

n

max

pXn,σLBB′n

1 n [I(X n; LBBn)τ − I(X n; E n)τ] , where τX nLBBnE n :=

xn pX n(xn)|xn

xn|X n ⊗ Mxn

B′n→BnE n(σLBB′n).

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 13 / 16

slide-45
SLIDE 45

Application: Private Reading E3 E2 ê

k

B3 B’3 B’2 B1

x1(k)

M

B’1 B2

A1

x2(k)

M

A2

x3(k)

M

L L L E1

B1 B2 B3

Private reading: Eve present when Bob performs the readout: Wiretap memory cell ¯ SX = {Mx

B′→BE}x∈X .

Special case: Isometric wiretap memory cell ¯ Siso

X .

The non-adaptive private reading capacity of a wiretap memory cell ¯ SX is given by Pread

n-a

¯

SX

  • = sup

n

max

pXn,σLBB′n

1 n [I(X n; LBBn)τ − I(X n; E n)τ] , where τX nLBBnE n :=

xn pX n(xn)|xn

xn|X n ⊗ Mxn

B′n→BnE n(σLBB′n).

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 13 / 16

slide-46
SLIDE 46

Private reading capacity

The strong converse private reading capacity Pread( ¯ Siso

X ) of an isometric wiretap memory cell

¯ Siso

X = {UMx B′→BE}x∈X is bounded from above as

  • Pread( ¯

Siso

X ) ≤ E 2→2 max (N ¯ S XB′→XB),

where N

¯ S XB′→XB(·) := TrE

  • U

¯ S XB′→XBE(·)

  • U

¯ S XB′→XBE

, such that U

¯ S XB′→XBE :=

  • x∈X

|x x|X ⊗ UMx

B′→BE.

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 14 / 16

slide-47
SLIDE 47

Conclusion

We derived upper bounds on entanglement generation and secret-key-agreement capacities over bidirectional channels. Sizes of reference systems are same as size of input systems (Open question in [BHLS03]). Obtain tighter upper bounds for channels obeying certain symmetries, see [DBW17]. Introduced secure protocol for reading of memory devices under scrutiny of an

  • eavesdropper. Both upper and lower bounds for this protocol can be found in [DBW17].

[DBW17] arXiv : 1712.00827

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 15 / 16

slide-48
SLIDE 48

Conclusion

We derived upper bounds on entanglement generation and secret-key-agreement capacities over bidirectional channels. Sizes of reference systems are same as size of input systems (Open question in [BHLS03]). Obtain tighter upper bounds for channels obeying certain symmetries, see [DBW17]. Introduced secure protocol for reading of memory devices under scrutiny of an

  • eavesdropper. Both upper and lower bounds for this protocol can be found in [DBW17].

[DBW17] arXiv : 1712.00827

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 15 / 16

slide-49
SLIDE 49

Conclusion

We derived upper bounds on entanglement generation and secret-key-agreement capacities over bidirectional channels. Sizes of reference systems are same as size of input systems (Open question in [BHLS03]). Obtain tighter upper bounds for channels obeying certain symmetries, see [DBW17]. Introduced secure protocol for reading of memory devices under scrutiny of an

  • eavesdropper. Both upper and lower bounds for this protocol can be found in [DBW17].

[DBW17] arXiv : 1712.00827

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 15 / 16

slide-50
SLIDE 50

Conclusion

We derived upper bounds on entanglement generation and secret-key-agreement capacities over bidirectional channels. Sizes of reference systems are same as size of input systems (Open question in [BHLS03]). Obtain tighter upper bounds for channels obeying certain symmetries, see [DBW17]. Introduced secure protocol for reading of memory devices under scrutiny of an

  • eavesdropper. Both upper and lower bounds for this protocol can be found in [DBW17].

[DBW17] arXiv : 1712.00827

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 15 / 16

slide-51
SLIDE 51

SWAP and Collective dephasing

0.2 0.4 0.6 0.8 1

p

1 1.2 1.4 1.6 1.8 2

R2-2

max

Collective Dephasing

2 2 /3 /2 2 /5 /3 2 /7 /4

Collective dephasing: |00 → |00, |01 → eiφ|01, |10 → eiφ|10, |11 → e2iφ|11. Swap operator S =

ij |ij

ji| and collective dephasing: NA′B′→AB(ρ) =pSρS† + (1 − p)UφSρS†Uφ†

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 16 / 16

slide-52
SLIDE 52

SWAP and Collective dephasing

0.2 0.4 0.6 0.8 1

p

1 1.2 1.4 1.6 1.8 2

R2-2

max

Collective Dephasing

2 2 /3 /2 2 /5 /3 2 /7 /4

Collective dephasing: |00 → |00, |01 → eiφ|01, |10 → eiφ|10, |11 → e2iφ|11. Swap operator S =

ij |ij

ji| and collective dephasing: NA′B′→AB(ρ) =pSρS† + (1 − p)UφSρS†Uφ†

Siddhartha Das (QST@LSU) Bipartite quantum interactions QCrypt ’18 16 / 16