Energy Condition, Modular Flow, and AdS/CFT Black Holes and - - PowerPoint PPT Presentation

energy condition modular flow and ads cft
SMART_READER_LITE
LIVE PREVIEW

Energy Condition, Modular Flow, and AdS/CFT Black Holes and - - PowerPoint PPT Presentation

Energy Condition, Modular Flow, and AdS/CFT Black Holes and Holography Workshop,TSIMF, Jan 7-11, 2019 Huajia Wang Kavli Institute for Theoretical Physics u v v ( z ) u X + B ( z ) v O z = 0 z arXiv:1806.10560; arXiv:1706.09432; JHEP


slide-1
SLIDE 1

Energy Condition, Modular Flow, and AdS/CFT

arXiv:1806.10560; arXiv:1706.09432; JHEP 1609 038(2016)

Huajia Wang Kavli Institute for Theoretical Physics

  • S. Balakrishnan, T. Faulkner, R. Leigh, M. Li, Z. Khandker, O. Parrikar, H. Wang

X+

B(z)

∆v(z)

z = 0

u

v

O

z

u

v

Black Holes and Holography Workshop,TSIMF, Jan 7-11, 2019

slide-2
SLIDE 2

Energy Conditions

unitarity of QM: positivity of total energy extended systems (QFT): local energy/momentum density constraints on energy/momentum density What are they?

E0

E ≥ 0

E = Z

Rn dxn E(x) ≥ 0

slide-3
SLIDE 3

Energy Conditions

Why do we care? classical: important in general relativity energy-momentum = spacetime geometry energy conditions = constraints on spacetime

Einstein’s equations:

Gµν = 8π Tµν

slide-4
SLIDE 4

Energy Conditions

Why do we care? strong energy condition (SEC) —> singularity theorem null energy condition (NEC) —> horizon area theorem examples:

Hawking, Ellis, 1973

✓ Tab − 1 2Tgab ◆ ζaζb ≥ 0

Tab kakb ≥ 0

x

t

k

x

t

ζ

slide-5
SLIDE 5

Energy Conditions

What do we want? in QM: QFTs in fixed background spacetime constraints on NEC violated by quantum effects: e.g. Casimir effect correct modification to NEC? two main conjectures: h ˆ Tµνiψ

Averaged Null Energy Condition (ANEC) Quantum Null Energy Condition (QNEC)

slide-6
SLIDE 6

AVERAGED NULL ENERGY CONDITION (ANEC)

R dλ h ˆ Tµνiψkµkν 0

x

t

k λ

affine parameter

Why? violation leads to causality breakdown

  • M. Morris, K. Thorne, U. Yurtsever, PRL. 61. 13. 1988

— supports traversable wormhole/time machine QUANTUM NULL ENERGY CONDITION (QNEC)

h ˆ Tµν(y)iψkµkν ∂2

λSA(λ)(ψ)

y

A(0)

λ

A(λ)

Motivation: generalized second law

slide-7
SLIDE 7

Can we prove them in QFTs? How?

slide-8
SLIDE 8

ANEC for free scalar and Maxwell fields; ANEC for 2d massive QFTs; QNEC for free/super-renormalizable fields;

  • G. Klinkhammer, 1991; L. Ford, T. Roman,

1995; A. Folacci, 1992

  • R. Verch, 2000
  • R. Busso, Z. Fisher, J. Koeller,
  • S. Leichenaber, A. Wall, 2015

for specific types of theories:

A brief history of proofs…

slide-9
SLIDE 9

A brief history of proofs…

for holographic theories (a broad class of CFTs): proof of ANEC using AdS/CFT: proof of QNEC using AdS/CFT:

  • W. Kelly, A. Wall, 2014

causality constraint in the bulk. entanglement wedge nesting (EWN)

  • J. Koeller, S. Leichenauer, 2016; C. Akers. V. Chandrasekaran,
  • S. Leichenaber, A. Levin, A. Moghaddam, 2017
slide-10
SLIDE 10

Can we do better? Proofs for generic QFT/CFTs?

slide-11
SLIDE 11

Can we do better? Proofs for generic QFT/CFTs?

Recent progresses…

slide-12
SLIDE 12

Can we do better? Proofs for generic QFT/CFTs?

ANEC in relativistic QFTs: ANEC in CFTs: QNEC in CFTs:

  • T. Faulkner, R. Leigh, O. Parrikar, H. Wang, 2016
  • T. Hartman, S. Kundu, A. Tajdini, 2016
  • S. Balakrishna, T. Faulkner, Z. Khandker, H. Wang, 2017

monotonicity of relative entropy causality of correlation functions in light-cone limit causality of correlation function under modular flow

slide-13
SLIDE 13

Plan of the talk: Review of AdS/CFT proofs (ANEC + QNEC) Summary of general field theory proofs (ANEC + QNEC) Bulk modular flow in AdS/CFT Conclusion/outlooks

slide-14
SLIDE 14

Plan of the talk: Review of AdS/CFT proofs (ANEC + QNEC) Summary of general field theory proofs (ANEC + QNEC) Bulk modular flow in AdS/CFT Conclusion/outlooks

slide-15
SLIDE 15

Proving ANEC using AdS/CFT

  • W. Kelly, A. Wall, 2014

Z ∞

−∞

dx+h ˆ T++iψ 0

slide-16
SLIDE 16

Proving ANEC using AdS/CFT

  • W. Kelly, A. Wall, 2014

Z ∞

−∞

dx+h ˆ T++iψ 0

“bulk respects boundary causality” t

x

z

Lbdry

Lbulk

Lbulk ≥ Lbdry

  • S. Gao, R. Wald, 2000
slide-17
SLIDE 17

Proving ANEC using AdS/CFT

  • W. Kelly, A. Wall, 2014

Z ∞

−∞

dx+h ˆ T++iψ 0

“bulk respects boundary causality” t

x

z

Lbdry

Lbulk

Lbulk ≥ Lbdry

  • S. Gao, R. Wald, 2000

As a GR result, can be proved by assuming that the “ANEC” in the bulk theory is satisfied

slide-18
SLIDE 18

Proving ANEC using AdS/CFT

  • W. Kelly, A. Wall, 2014

Z ∞

−∞

dx+h ˆ T++iψ 0

In AdS/CFT, via Fefferman-Graham gauge expansion:

ds2 = R2 z2 ⇢ dz2 +  ηab + zd 16πG dRd−1 hTabiψ + O

  • zd+2

dxadxb

  • , z ! 0

“bulk respects boundary causality” t

x

z

Lbdry

Lbulk

Lbulk ≥ Lbdry

  • S. Gao, R. Wald, 2000
slide-19
SLIDE 19

Proving ANEC using AdS/CFT

  • W. Kelly, A. Wall, 2014

Z ∞

−∞

dx+h ˆ T++iψ 0

Gao-Wald’s conclusion as consistent condition for holographic CFTs

lim z → 0

Z ∞

−∞

dx+h ˆ T++iψ 0

boundary ANEC

leading order constraint in F. G. gauge expansion

“bulk respects boundary causality” t

x

z

Lbdry

Lbulk

Lbulk ≥ Lbdry

  • S. Gao, R. Wald, 2000
slide-20
SLIDE 20
  • J. Koeller, S. Leichenauer, 2016; C. Akers. V. Chandrasekaran,
  • S. Leichenaber, A. Levin, A. Moghaddam, 2017

Proving QNEC using AdS/CFT

hTuuiψ ∂2

uSEE

slide-21
SLIDE 21
  • J. Koeller, S. Leichenauer, 2016; C. Akers. V. Chandrasekaran,
  • S. Leichenaber, A. Levin, A. Moghaddam, 2017

Proving QNEC using AdS/CFT

hTuuiψ ∂2

uSEE

“bulk reconstruction in entanglement wedges”

z

t

x

A

D(A)

how much bulk region can be reconstructed from CFT

  • perators localized in D(A)?

AdS/CFT: bulk physics can be “reconstructed” from the boundary subregion duality

slide-22
SLIDE 22
  • J. Koeller, S. Leichenauer, 2016; C. Akers. V. Chandrasekaran,
  • S. Leichenaber, A. Levin, A. Moghaddam, 2017

Proving QNEC using AdS/CFT

hTuuiψ ∂2

uSEE

“bulk reconstruction in entanglement wedges”

z

t

x

A

D(A)

a

RT surface Σ strong evidence: entanglement wedge

  • X. Dong, D. Harlow, A. Wall, 2016

∂a = Σ ∪ A

slide-23
SLIDE 23
  • J. Koeller, S. Leichenauer, 2016; C. Akers. V. Chandrasekaran,
  • S. Leichenaber, A. Levin, A. Moghaddam, 2017

Proving QNEC using AdS/CFT

hTuuiψ ∂2

uSEE

“bulk reconstruction in entanglement wedges”

z

t

x

A

D(A)

a

RT surface Σ strong evidence: entanglement wedge

  • X. Dong, D. Harlow, A. Wall, 2016

∂a = Σ ∪ A

entanglement wedge = D(a)

D(a)

slide-24
SLIDE 24
  • J. Koeller, S. Leichenauer, 2016; C. Akers. V. Chandrasekaran,
  • S. Leichenaber, A. Levin, A. Moghaddam, 2017

Proving QNEC using AdS/CFT

hTuuiψ ∂2

uSEE

“bulk reconstruction in entanglement wedges”

z

t

x

A

D(A)

strong evidence: entanglement wedge

  • X. Dong, D. Harlow, A. Wall, 2016

a

∂a = Σ ∪ A

RT surface Σ entanglement wedge = D(a)

D(a)“ ≈ ”D(A)

D(a)

slide-25
SLIDE 25
  • J. Koeller, S. Leichenauer, 2016; C. Akers. V. Chandrasekaran,
  • S. Leichenaber, A. Levin, A. Moghaddam, 2017

Proving QNEC using AdS/CFT

hTuuiψ ∂2

uSEE

Entanglement Wedge Nesting (EWN):

D( ˜ A) ⊆ D(A) → D(˜ a) ⊆ D(a)

slide-26
SLIDE 26
  • J. Koeller, S. Leichenauer, 2016; C. Akers. V. Chandrasekaran,
  • S. Leichenaber, A. Levin, A. Moghaddam, 2017

Proving QNEC using AdS/CFT

hTuuiψ ∂2

uSEE

Entanglement Wedge Nesting (EWN):

D( ˜ A) ⊆ D(A) → D(˜ a) ⊆ D(a)

z

A

˜ A

u

v

∆u

at the boundary: D( ˜ A) ⊆ D(A) null deformation

∆u ≥ 0 :

slide-27
SLIDE 27
  • J. Koeller, S. Leichenauer, 2016; C. Akers. V. Chandrasekaran,
  • S. Leichenaber, A. Levin, A. Moghaddam, 2017

Proving QNEC using AdS/CFT

hTuuiψ ∂2

uSEE

Entanglement Wedge Nesting (EWN):

D( ˜ A) ⊆ D(A) → D(˜ a) ⊆ D(a)

z

A

˜ A

u

v

∆u

a ˜ a

ΣA Σ ˜

A

into the bulk: at the boundary: D( ˜ A) ⊆ D(A) null deformation

∆u ≥ 0 :

(EWN)

D(˜ a) ⊆ D(a)

Σ ˜

A

ΣA

spacelike/null

RT surfaces dynamics

slide-28
SLIDE 28
  • J. Koeller, S. Leichenauer, 2016; C. Akers. V. Chandrasekaran,
  • S. Leichenaber, A. Levin, A. Moghaddam, 2017

Proving QNEC using AdS/CFT

hTuuiψ ∂2

uSEE

Entanglement Wedge Nesting (EWN):

D( ˜ A) ⊆ D(A) → D(˜ a) ⊆ D(a)

Σ ˜

A

ΣA

spacelike/null

near boundary expansion: (F-G gauge)

guu = 16πG dRd−3 zd−2hTabiψ + O(zd)

Xi

Σ ˜

A(z) = Xi

∂ ˜ A +

4G dRd−1 zd∂iSEE( ˜ A) + O(zd+1)

Xi

ΣA(z) = Xi ∂A +

4G dRd−1 zd∂iSEE(A) + O(zd+1)

z

A

˜ A

u

v

∆u

a ˜ a

ΣA Σ ˜

A

slide-29
SLIDE 29
  • J. Koeller, S. Leichenauer, 2016; C. Akers. V. Chandrasekaran,
  • S. Leichenaber, A. Levin, A. Moghaddam, 2017

Proving QNEC using AdS/CFT

hTuuiψ ∂2

uSEE

Entanglement Wedge Nesting (EWN):

D( ˜ A) ⊆ D(A) → D(˜ a) ⊆ D(a)

Σ ˜

A

ΣA

spacelike/null

z

A

˜ A

u

v

∆u

a ˜ a

ΣA Σ ˜

A

z → 0

∆u → 0

hTuuiψ ∂2

uSEE 0

boundary QNEC

slide-30
SLIDE 30

Plan of the talk: Review of AdS/CFT proofs (ANEC + QNEC) Summary of general field theory proofs (ANEC + QNEC) Bulk modular flow in AdS/CFT Conclusion/outlooks

slide-31
SLIDE 31

Proving ANEC in relativistic QFTs

Z ∞

−∞

dx+h ˆ T++iψ 0

  • T. Faulkner, R. Leigh, O. Parrikar, H. Wang, 2016
slide-32
SLIDE 32

Proving ANEC in relativistic QFTs

Z ∞

−∞

dx+h ˆ T++iψ 0

  • T. Faulkner, R. Leigh, O. Parrikar, H. Wang, 2016

difficult using conventional QFT techniques surprising origin in information theory manifested by probing the entanglement structure

slide-33
SLIDE 33

Proving ANEC in relativistic QFTs

Z ∞

−∞

dx+h ˆ T++iψ 0

  • T. Faulkner, R. Leigh, O. Parrikar, H. Wang, 2016

Modular Hamiltonian:

A Ac

A = − ln ρΨ A ⊗ 1Ac + 1A ⊗ ln ρΨ Ac = HΨ A − HΨ Ac

A : Hfull → Hfull

A|Ψi = 0

slide-34
SLIDE 34

Proving ANEC in relativistic QFTs

Z ∞

−∞

dx+h ˆ T++iψ 0

  • T. Faulkner, R. Leigh, O. Parrikar, H. Wang, 2016

Modular Hamiltonian:

A Ac

encodes more detailed entanglement data in general, complicated and non-local simplifies in special cases

A = − ln ρΨ A ⊗ 1Ac + 1A ⊗ ln ρΨ Ac = HΨ A − HΨ Ac

A : Hfull → Hfull

A|Ψi = 0

Ψ = |vaci, A = half-space, KΨ

A = 2π

Z dd−1x x1T00 = Rindler Hamiltonian

e.g.

slide-35
SLIDE 35

Monotonicity property:

D( ˜ A) ⊆ D(A)

A Ac ˜ A

x+

x−

~ y

˜ A = A + ~ ⇠(y)

Proving ANEC in relativistic QFTs

Z ∞

−∞

dx+h ˆ T++iψ 0

  • T. Faulkner, R. Leigh, O. Parrikar, H. Wang, 2016

ψ is arbitrary

where

hKvac

˜ A iψ  hKvac A iψ

slide-36
SLIDE 36

Monotonicity property:

D( ˜ A) ⊆ D(A)

A Ac ˜ A

x+

x−

~ y

˜ A = A + ~ ⇠(y)

Proving ANEC in relativistic QFTs

Z ∞

−∞

dx+h ˆ T++iψ 0

  • T. Faulkner, R. Leigh, O. Parrikar, H. Wang, 2016

Why? Monotonicity of relative entropy SA(ψ|φ) = trρA(ψ) ln [ρA(ψ)/ρA(φ)] measure of “distinguishability”

S ˜

A(ψ|φ) ≤ SA(ψ|φ)

D( ˜ A) ⊆ D(A) for

|φi = |vaci

for special case of : hKvac

˜ A iψ  hKvac A iψ

ψ is arbitrary

where

hKvac

˜ A iψ  hKvac A iψ

slide-37
SLIDE 37

Monotonicity property:

D( ˜ A) ⊆ D(A)

A Ac ˜ A

x+

x−

~ y

˜ A = A + ~ ⇠(y)

Proving ANEC in relativistic QFTs

Z ∞

−∞

dx+h ˆ T++iψ 0

  • T. Faulkner, R. Leigh, O. Parrikar, H. Wang, 2016

ψ is arbitrary

where

hKvac

˜ A iψ  hKvac A iψ

perturbation theory: half-space,

A = Kvac

A

= Rindler Hamiltonian hKvac

˜ A iψ  hKvac A iψ

⇠+(~ y) > 0

for arbitrary null requiring ANEC

“=”

slide-38
SLIDE 38

Proving QNEC in general CFTs

hTuuiψ ∂2

uSEE

  • S. Balakrishna, T. Faulkner, Z. Khandker, H. Wang, 2017
slide-39
SLIDE 39

ANEC proof from entanglement structure alternative proof of ANEC from causality of correlation function combine entanglement structure + causality? proof of QNEC (stronger conjecture)!

Proving QNEC in general CFTs

hTuuiψ ∂2

uSEE

  • S. Balakrishna, T. Faulkner, Z. Khandker, H. Wang, 2017
  • T. Hartman, S. Kundu, A. Tajdini, 2016
slide-40
SLIDE 40

Proving QNEC in general CFTs

hTuuiψ ∂2

uSEE

  • S. Balakrishna, T. Faulkner, Z. Khandker, H. Wang, 2017

f(u, v) / hψ|O(u, v)O(u, v)|ψi

causality of correlation function:

slide-41
SLIDE 41

Proving QNEC in general CFTs

hTuuiψ ∂2

uSEE

  • S. Balakrishna, T. Faulkner, Z. Khandker, H. Wang, 2017

ψ

ψ†

u v

O

O

f(u, v) / hψ|O(u, v)O(u, v)|ψi

hψ| [O, O] |ψi = 0

Causality:

uv < 0

for

causality of correlation function:

slide-42
SLIDE 42

Proving QNEC in general CFTs

hTuuiψ ∂2

uSEE

  • S. Balakrishna, T. Faulkner, Z. Khandker, H. Wang, 2017

ψ

ψ†

u v

O

O

f(u, v) / hψ|O(u, v)O(u, v)|ψi

hψ| [O, O] |ψi = 0

Causality:

uv < 0

for

causality of correlation function: “dress” the correlator to probe entanglement structure? modular flow: O → OA(s) ≡ eis Kψ

AOe−is Kψ A

in general: highly non-local!

slide-43
SLIDE 43

Proving QNEC in general CFTs

hTuuiψ ∂2

uSEE

  • S. Balakrishna, T. Faulkner, Z. Khandker, H. Wang, 2017

˜ A A Ac ˜ Ac

u v

δu

consider:

O1 O2

f(s) = N −1hψ|O

˜ A 1 (s)OA 2 (s)|ψi

O

˜ A 1 (s) = eis Kψ

˜ AO1e−is Kψ ˜ A

OA

2 (s) = eis Kψ

AO2e−is Kψ A

slide-44
SLIDE 44

Proving QNEC in general CFTs

hTuuiψ ∂2

uSEE

  • S. Balakrishna, T. Faulkner, Z. Khandker, H. Wang, 2017

˜ A A Ac ˜ Ac

u v

δu

consider:

O1 O2

Tomita-Takesaki theory (in algebraic QFT):

O ∈ MA → OA(s) ∈ MA, s ∈ R

: von Neumann algebra associated with A, i.e. operators supported in D(A)

MA

f(s) = N −1hψ|O

˜ A 1 (s)OA 2 (s)|ψi

O

˜ A 1 (s) = eis Kψ

˜ AO1e−is Kψ ˜ A

OA

2 (s) = eis Kψ

AO2e−is Kψ A

slide-45
SLIDE 45

Proving QNEC in general CFTs

hTuuiψ ∂2

uSEE

  • S. Balakrishna, T. Faulkner, Z. Khandker, H. Wang, 2017

˜ A A Ac ˜ Ac

u v

δu

consider:

O1 O2

Tomita-Takesaki theory (in algebraic QFT):

O

˜ A 1 (s)

O

˜ A 1 (s)

D( ˜ A)

is supported only in

f(s) = N −1hψ|O

˜ A 1 (s)OA 2 (s)|ψi

O

˜ A 1 (s) = eis Kψ

˜ AO1e−is Kψ ˜ A

OA

2 (s) = eis Kψ

AO2e−is Kψ A

D(Ac)

is supported only in

OA

2 (s)

OA

2 (s)

slide-46
SLIDE 46

Proving QNEC in general CFTs

hTuuiψ ∂2

uSEE

  • S. Balakrishna, T. Faulkner, Z. Khandker, H. Wang, 2017

˜ A A Ac ˜ Ac

u v

δu

consider:

O1 O2

Tomita-Takesaki theory (in algebraic QFT):

O

˜ A 1 (s)

[ ]

O

˜ A 1 (s) ,

= 0

s ∈ R

for

a subtler notion of causality: hidden in entanglement structure!

OA

2 (s)

f(s) = N −1hψ|O

˜ A 1 (s)OA 2 (s)|ψi

O

˜ A 1 (s) = eis Kψ

˜ AO1e−is Kψ ˜ A

OA

2 (s) = eis Kψ

AO2e−is Kψ A

OA

2 (s)

slide-47
SLIDE 47

Proving QNEC in general CFTs

hTuuiψ ∂2

uSEE

  • S. Balakrishna, T. Faulkner, Z. Khandker, H. Wang, 2017

Outline of the proof:

  • 1. Unitarity + Cauchy-Schwarz inequality:

Ref(s) ≤ 1, Im s = ±π/2

˜ A A Ac ˜ Ac

u

v

δu

O1

O2

slide-48
SLIDE 48

Proving QNEC in general CFTs

hTuuiψ ∂2

uSEE

  • S. Balakrishna, T. Faulkner, Z. Khandker, H. Wang, 2017

Outline of the proof:

δu

  • 2. Causality: analytic continuation of f(s)

into the complex stripe {−π < Im s < π}

  • 1. Unitarity + Cauchy-Schwarz inequality:

Ref(s) ≤ 1, Im s = ±π/2

˜ A A Ac ˜ Ac

u

v

δu

O1

O2

slide-49
SLIDE 49

Proving QNEC in general CFTs

hTuuiψ ∂2

uSEE

  • S. Balakrishna, T. Faulkner, Z. Khandker, H. Wang, 2017

Outline of the proof:

˜ A A Ac ˜ Ac

u

v

δu

  • 2. Causality: analytic continuation of f(s)

into the complex stripe {−π < Im s < π}

  • 3. Light-cone limit expansion:

f(s) = 1 + C−1

T esu(−uv)

d−2 2 IQ + ...

v → 0, u fixed

  • 1. Unitarity + Cauchy-Schwarz inequality:

Ref(s) ≤ 1, Im s = ±π/2

IQ = Z δu du0Tuu(u0) + δSEE(A) δu − δSEE( ˜ A) δu !

O1

O2

slide-50
SLIDE 50

Proving QNEC in general CFTs

hTuuiψ ∂2

uSEE

  • S. Balakrishna, T. Faulkner, Z. Khandker, H. Wang, 2017

Outline of the proof:

˜ A A Ac ˜ Ac

u

v

δu

  • 4. derive a sum rule (using the analytic

continuation) + unitarity bound:

IQ ∝ Z

Im s=±π/2

ds [1 − Ref(s)] ≥ 0

O1

O2

slide-51
SLIDE 51

Proving QNEC in general CFTs

hTuuiψ ∂2

uSEE

  • S. Balakrishna, T. Faulkner, Z. Khandker, H. Wang, 2017

Outline of the proof:

˜ A A Ac ˜ Ac

u

v

δu

  • 4. derive a sum rule (using the analytic

continuation) + unitarity bound:

IQ ∝ Z

Im s=±π/2

ds [1 − Ref(s)] ≥ 0

IQ = Z δu du0Tuu(u0) + δSEE(A) δu δSEE( ˜ A) δu ! ⇡ δu

  • hTuuiψ ∂2

uSEE

  • (lim δu ! 0) ! hTuuiψ ∂2

uSEE

QNEC

O1

O2

slide-52
SLIDE 52

Plan of the talk: Review of AdS/CFT proofs (ANEC + QNEC) Summary of general field theory proofs (ANEC + QNEC) Bulk modular flow in AdS/CFT Conclusion/outlooks

slide-53
SLIDE 53

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

in holography, EWN near boundary = boundary QNEC

slide-54
SLIDE 54

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

in holography, EWN near boundary = boundary QNEC “universality” of holography near boundary

slide-55
SLIDE 55

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

in holography, EWN near boundary = boundary QNEC “universality” of holography near boundary reproducible in generic CFTs (not necessarily holographic)

slide-56
SLIDE 56

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

in holography, EWN near boundary = boundary QNEC “universality” of holography near boundary reproducible in generic CFTs (not necessarily holographic) the CFT proof: EWN (near boundary) in disguise

slide-57
SLIDE 57

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

in holography, EWN near boundary = boundary QNEC “universality” of holography near boundary reproducible in generic CFTs (not necessarily holographic) the CFT proof: EWN (near boundary) in disguise modular flow in the boundary “knows” RT surface dynamics

e.g. T. Faulkner, A.Lewkowycz, 2017

slide-58
SLIDE 58

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

in holography, EWN near boundary = boundary QNEC “universality” of holography near boundary reproducible in generic CFTs (not necessarily holographic) the CFT proof: EWN (near boundary) in disguise modular flow in the boundary “knows” RT surface dynamics understand this connection more explicitly

e.g. T. Faulkner, A.Lewkowycz, 2017

slide-59
SLIDE 59

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

in holography, EWN near boundary = boundary QNEC “universality” of holography near boundary reproducible in generic CFTs (not necessarily holographic) the CFT proof: EWN (near boundary) in disguise modular flow in the boundary “knows” RT surface dynamics understand this connection more explicitly a concrete step: bulk approach for computing f(s)

e.g. T. Faulkner, A.Lewkowycz, 2017

slide-60
SLIDE 60

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

Revisit

˜ A A Ac ˜ Ac

u

v

δu

O1 O2

f(s) / hψ|O

˜ A 1 (s)OA 2 (s)|ψi

slide-61
SLIDE 61

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

from “Heisenberg” to “Schrodinger" picture:

˜ A A Ac ˜ Ac

u

v

δu

O1 O2

Revisit

f(s) / hψ|O

˜ A 1 (s)OA 2 (s)|ψi

f(s) / hψ|eisKψ

˜ AO1e−isKψ ˜ A eisKψ AO2e−isKψ A|ψi

slide-62
SLIDE 62

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

from “Heisenberg” to “Schrodinger" picture:

˜ A A Ac ˜ Ac

u

v

δu

O1 O2

A = Hψ A ⊗ 1Ac − 1A ⊗ Hψ Ac

recall

A,Ac = half-sided modular Hamiltonian

,

f(s) / hψ|eisKψ

˜ AO1e−isKψ ˜ A eisKψ AO2e−isKψ A|ψi

⌘ hψ|eisHψ

˜ A−isHψ ˜ Ac O1e−isHψ ˜ A+isHψ ˜ Ac eisHψ A−isHψ Ac O2e−isHψ A+isHψ Ac |ψi

Revisit

f(s) / hψ|O

˜ A 1 (s)OA 2 (s)|ψi

slide-63
SLIDE 63

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

from “Heisenberg” to “Schrodinger" picture:

˜ A A Ac ˜ Ac

u

v

δu

O1 O2

A = Hψ A ⊗ 1Ac − 1A ⊗ Hψ Ac

recall

A,Ac = half-sided modular Hamiltonian

,

f(s) / hψ|eisKψ

˜ AO1e−isKψ ˜ A eisKψ AO2e−isKψ A|ψi

⌘ hψ|eisHψ

˜ A−isHψ ˜ Ac O1e−isHψ ˜ A+isHψ ˜ Ac eisHψ A−isHψ Ac O2e−isHψ A+isHψ Ac |ψi

= hψ|e−isHψ

Ac+isHψ ˜ A O1 O2 e−isHψ ˜ A+isHψ Ac |ψi

Revisit

f(s) / hψ|O

˜ A 1 (s)OA 2 (s)|ψi

[Hψ

Ac, ˜ Ac, O1] = 0, [Hψ A, ˜ A, O2] = 0

using

slide-64
SLIDE 64

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

from “Heisenberg” to “Schrodinger" picture:

˜ A A Ac ˜ Ac

u

v

δu

O1 O2

A = Hψ A ⊗ 1Ac − 1A ⊗ Hψ Ac

recall

A,Ac = half-sided modular Hamiltonian

,

f(s) / hψ|eisKψ

˜ AO1e−isKψ ˜ A eisKψ AO2e−isKψ A|ψi

⌘ hψ|eisHψ

˜ A−isHψ ˜ Ac O1e−isHψ ˜ A+isHψ ˜ Ac eisHψ A−isHψ Ac O2e−isHψ A+isHψ Ac |ψi

= hψ|e−isHψ

Ac+isHψ ˜ A O1 O2 e−isHψ ˜ A+isHψ Ac |ψi

Revisit

f(s) / hψ|O

˜ A 1 (s)OA 2 (s)|ψi

[Hψ

Ac, ˜ Ac, O1] = 0, [Hψ A, ˜ A, O2] = 0

using

= hψ|e−isHψ

AeisHψ ˜ AO1 O2e−isHψ ˜ AeisHψ A|ψi

A|ψi = 0 ! Hψ A|ψi = Hψ Ac|ψi

recall etc

slide-65
SLIDE 65

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

from “Heisenberg” to “Schrodinger" picture:

˜ A A Ac ˜ Ac

u

v

δu

O1 O2

A = Hψ A ⊗ 1Ac − 1A ⊗ Hψ Ac

recall

A,Ac = half-sided modular Hamiltonian

,

⌘ hψ|eisHψ

˜ A−isHψ ˜ Ac O1e−isHψ ˜ A+isHψ ˜ Ac eisHψ A−isHψ Ac O2e−isHψ A+isHψ Ac |ψi

= hψ|e−isHψ

Ac+isHψ ˜ A O1 O2 e−isHψ ˜ A+isHψ Ac |ψi

Revisit

f(s) / hψ|O

˜ A 1 (s)OA 2 (s)|ψi

[Hψ

Ac, ˜ Ac, O1] = 0, [Hψ A, ˜ A, O2] = 0

using

= hψ|e−isHψ

AeisHψ ˜ AO1 O2e−isHψ ˜ AeisHψ A|ψi

A|ψi = 0 ! Hψ A|ψi = Hψ Ac|ψi

recall etc

= hψA, ˜

A(s)|O1 O2|ψA, ˜ A(s)i

|ψA, ˜

A(s)i = e−isHψ

˜ AeisHψ A|ψi

where

f(s) / hψ|eisKψ

˜ AO1e−isKψ ˜ A eisKψ AO2e−isKψ A|ψi

slide-66
SLIDE 66

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

˜ A A Ac ˜ Ac

u

v

δu

O1 O2

So, in “Schrodinger" picture:

|ψA, ˜

A(s)i = e−isHψ

˜ AeisHψ A|ψi

where

f(s) / hψA, ˜

A(s)|O1O2|ψA, ˜ A(s)i

slide-67
SLIDE 67

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

˜ A A Ac ˜ Ac

u

v

δu

O1 O2

So, in “Schrodinger" picture:

|ψA, ˜

A(s)i = e−isHψ

˜ AeisHψ A|ψi

where

f(s) / hψA, ˜

A(s)|O1O2|ψA, ˜ A(s)i

to use AdS/CFT, consider: in a holographic CFT bulk dual of has smooth geometry conformal dimension of :

|ψi

O1,2

1 ⌧ ∆ ⌧ `AdS/`plank

slide-68
SLIDE 68

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

˜ A A Ac ˜ Ac

u

v

δu

O1 O2

So, in “Schrodinger" picture:

|ψA, ˜

A(s)i = e−isHψ

˜ AeisHψ A|ψi

where

f(s) / hψA, ˜

A(s)|O1O2|ψA, ˜ A(s)i

Geodesic approximation:

hψ| |ψi

O1O2

t

x

z O1 O2

|ψi

L(x1, x2)

≈ exp [−mL(x1, x2)]

slide-69
SLIDE 69

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

˜ A A Ac ˜ Ac

u

v

δu

O1 O2

So, in “Schrodinger" picture:

|ψA, ˜

A(s)i = e−isHψ

˜ AeisHψ A|ψi

where

f(s) / hψA, ˜

A(s)|O1O2|ψA, ˜ A(s)i

Geodesic approximation:

O1 O2

hψA, ˜

A(s)|

|ψA, ˜

A(s)i

t

x

z O1 O2

|ψA, ˜

A(s)i

Ls

A, ˜ A(x1, x2)

≈ exp h −mLs

A, ˜ A(x1, x2)

i

slide-70
SLIDE 70

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

consider a simpler case:

t x

z

|ψi

t

x

z

How?

|ψA(s)i

|ψA(s)i = eisHψ

A|ψi

i.e. “single modular flow”

slide-71
SLIDE 71

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

consider a simpler case:

t x

z

|ψi

t

x

z

How?

|ψA(s)i

|ψA(s)i = eisHψ

A|ψi

i.e. “single modular flow”

hψA(s)|OA|ψA(s)i = hψ|e−isHψ

AOAeisHψ A|ψi = hψ|e−isHψ Ac OAeisHψ Ac |ψi

= hψ|e−isHψ

Ac eisHψ Ac OA|ψi = hψ|OA|ψi

OA

D(A)

for any supported only in hint:

hOAiψA(s) = hOAiψ

:

slide-72
SLIDE 72

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

consider a simpler case:

t x

z

|ψi

t

x

z

How?

|ψA(s)i

|ψA(s)i = eisHψ

A|ψi

i.e. “single modular flow”

OA

D(A)

for any supported only in hint:

hOAiψA(s) = hOAiψ

:

similarly, for any supported only in

:

OAc

D(Ac)

hOAciψA(s) = hOAciψ

slide-73
SLIDE 73

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

entanglement wedge reconstruction:

D(a)“ ≈ ”D(A), D(ac)“ ≈ ”D(Ac)

|ψA(s)i

a

ac

slide-74
SLIDE 74

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

entanglement wedge reconstruction:

D(a)“ ≈ ”D(A), D(ac)“ ≈ ”D(Ac)

a

ac

ΣA

in

:

same metric same bulk fields, etc

00ψA(s) ≡ ψ”

e.g. entanglement wedges

|ψA(s)i

slide-75
SLIDE 75

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

entanglement wedge reconstruction:

D(a)“ ≈ ”D(A), D(ac)“ ≈ ”D(Ac)

a

ac

ΣA

in

:

00ψA(s) ≡ ψ”

entanglement wedges

|ψA(s)i

? ?

in

:

“Milne” wedges unknown, possibly with kinks/singularities

slide-76
SLIDE 76

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

a

ΣA

? ?

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

entanglement wedge reconstruction:

D(a)“ ≈ ”D(A), D(ac)“ ≈ ”D(Ac)

ac

|ψA(s)i

geodesic: a function

  • f {x1, x2, s}

generic geodesics pass through both the entanglement and “Milne” wedges we don’t know what to do…

slide-77
SLIDE 77

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

a

ΣA

? ?

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

entanglement wedge reconstruction:

D(a)“ ≈ ”D(A), D(ac)“ ≈ ”D(Ac)

ac

|ψA(s)i

geodesic: a function

  • f {x1, x2, s}

if we fine-tune one

  • f the parameters:

e.g. s = s(x1, x2) the geodesic avoids the Milne wedge, passes through ΣA

slide-78
SLIDE 78

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

So, what do we know about geodesics in the entanglement wedges (EW)?

ΣA

|ψA(s)i

La Lac

each segment is a geodesic in the

  • riginal geometry

{La, Lac}

|ψi u v

slide-79
SLIDE 79

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

So, what do we know about geodesics in the entanglement wedges (EW)?

ΣA

|ψA(s)i

La Lac

each segment is a geodesic in the

  • riginal geometry .

modular flow affects the matching condition at .

{La, Lac}

|ψi

ΣA

u v

slide-80
SLIDE 80

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

So, what do we know about geodesics in the entanglement wedges (EW)?

ΣA

|ψA(s)i

La Lac

each segment is a geodesic in the

  • riginal geometry .

modular flow affects the matching condition at . JLMS (2015): .

{La, Lac}

|ψi

ΣA

A(bdry) =

ˆ A 4G + Hψ

a (bulk)

u v

slide-81
SLIDE 81

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

So, what do we know about geodesics in the entanglement wedges (EW)?

ΣA

|ψA(s)i

La Lac

each segment is a geodesic in the

  • riginal geometry .

modular flow affects the matching condition at . JLMS (2015): . is a constant in EW, .

{La, Lac}

|ψi

ΣA

A(bdry) =

ˆ A 4G + Hψ

a (bulk)

ˆ A eisHψ

A(bdry) ∝ eisHψ a (bulk)

u v

slide-82
SLIDE 82

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

So, what do we know about geodesics in the entanglement wedges (EW)?

ΣA

|ψA(s)i

La Lac

each segment is a geodesic in the

  • riginal geometry .

modular flow affects the matching condition at . JLMS (2015): . is a constant in EW, . bulk theory free (leading ordering 1/N): close to , acts like bulk Rindler Hamiltonian and generates boosts.

{La, Lac}

|ψi

ΣA

A(bdry) =

ˆ A 4G + Hψ

a (bulk)

ˆ A eisHψ

A(bdry) ∝ eisHψ a (bulk)

ΣA

a (bulk)

u v

slide-83
SLIDE 83

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

So, what do we know about geodesics in the entanglement wedges (EW)? matching condition: relative boost

  • f rapidity across .

s

ΣA

ΣA

|ψA(s)i

La Lac s

u v

slide-84
SLIDE 84

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

So, what do we know about geodesics in the entanglement wedges (EW)? matching condition: relative boost

  • f rapidity across .

s

ΣA modified notion of smoothness for curves across in . ΣA |ψA(s)i

ΣA

|ψA(s)i

La Lac s

u v

slide-85
SLIDE 85

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

So, what do we know about geodesics in the entanglement wedges (EW)? matching condition: relative boost

  • f rapidity across .

s

ΣA modified notion of smoothness for curves across in . ΣA |ψA(s)i

ξ ∈ ΣA

ΣA

|ψA(s)i

La Lac s

pk [L(ξ, x1)] = pk [L(ξ, x2)]

fine-tuning: identify s.t. at ξ

s(x1, x2) = 1 4π ln ✓pu [L(ξ, x1)] pv [L(ξ, x1)] ◆ ✓pv [L(ξ, x2)] pu [L(ξ, x2)] ◆

then

u v

slide-86
SLIDE 86

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

So, what do we know about geodesics in the entanglement wedges (EW)?

|ψA(s)i

La Lac s

u v

Therefore, for s∗ = s(x1, x2)

hO1OA

2 (s∗)iψ = hO1O2iψA(s∗)

≈ exp [−mL(ξ, x1) − mL(ξ, x2)]

ξ ∈ ΣA x1 x2

slide-87
SLIDE 87

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

We can extend this to the “double modular flow”: |ψA, ˜

A(s)i = e−isHψ

˜ AeisHψ A|ψi

x2

x1

ξA ∈ ΣA

ξ ˜

A ∈ Σ ˜ A

u v

|ψA, ˜

A(s)i

ξA ∈ ΣA, ξ ˜

A ∈ Σ ˜ A

matching conditions at select s∗ = s(x1, x2) hOA

1 (s∗)O ˜ A 2 (s∗)iψ = hO1O2iψA, ˜

A(s∗)

≈ exp [−m (L(ξA, x1) + L(ξ ˜

A, ξA) + L(ξ ˜ A, x2))]

slide-88
SLIDE 88

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

We can extend this to the “double modular flow”: |ψA, ˜

A(s)i = e−isHψ

˜ AeisHψ A|ψi

x2

x1

ξA ∈ ΣA

ξ ˜

A ∈ Σ ˜ A

u v

|ψA, ˜

A(s)i

ξA ∈ ΣA, ξ ˜

A ∈ Σ ˜ A

matching conditions at select s∗ = s(x1, x2) in the near boundary limit , successfully reproduced the CFT result in the light-cone limit

z → 0

z ∝ uv

slide-89
SLIDE 89

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

Applications: “single modular flow” with s = iπ Mirror conjugation:

OJ = eπKψAOe−πKψA = OA(iπ)

  • K. Papadodimas, S. Raju, 2014

iπ boost = reflection

x

y

z

O1

Ac A

ΣA

ξA

hOJ

1 O1iψ ⇡ exp [2mL(ξA, x1)]

fπ / hOA

1 (iπ)O1iψ

slide-90
SLIDE 90

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

Applications: “single modular flow” with s = iπ Mirror conjugation:

OJ = eπKψAOe−πKψA = OA(iπ)

  • K. Papadodimas, S. Raju, 2014

iπ boost = reflection

x

y

z

O1

Ac A

ΣA

ξA

hOJ

1 O1iψ ⇡ exp [2mL(ξA, x1)]

“OJ

1 ”

RT surface serves as a mirror for implementing conjugation

fπ / hOA

1 (iπ)O1iψ

slide-91
SLIDE 91

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

Applications: entanglement wedge nesting (EWN)

f(s) / hO

˜ A 1 (s + iπ)OA 1 (s)iψ

consider:

˜ A = A + δA

,

δA → 0

f(s) = hOJ

1 O1iψ ⇡ exp [2mL(ξA, x1)] for all s

, for

z

A ˜ A

O1

ΣA

Σ ˜

A

ξA

ξ ˜

A

m−1 ln " hO ˜

A 1 (s + iπ)OA 1 (s)iψ

hOA

1 (iπ)O1iψ

#

≈ L (ξ ˜

A, ξA) + O(δA2)

slide-92
SLIDE 92

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

Applications: entanglement wedge nesting (EWN)

f(s) / hO

˜ A 1 (s + iπ)OA 1 (s)iψ

consider:

˜ A = A + δA

,

δA → 0

f(s) = hOJ

1 O1iψ ⇡ exp [2mL(ξA, x1)] for all s

, for

max ( ln " hO ˜

A 1 (s + iπ)OA 1 (s)iψ

hOA

1 (iπ)O1iψ

# , s 2 R )  0

z

A ˜ A

O1

ΣA

Σ ˜

A

ξA

ξ ˜

A

EWN in CFT ( for ) |δA| ⌧ |A|

δA

for space-like

slide-93
SLIDE 93

Bulk modular flow in AdS/CFT

  • T. Faulkner, M. Li, H. Wang, 2018

Applications: entanglement wedge nesting (EWN)

f(s) / hO

˜ A 1 (s + iπ)OA 1 (s)iψ

consider:

˜ A = A + δA

,

δA → 0

f(s) = hOJ

1 O1iψ ⇡ exp [2mL(ξA, x1)] for all s

, for

z

A ˜ A

O1

ΣA

Σ ˜

A

ξA

ξ ˜

A

in Tomita-Takaseki theory:

|U(t)| ≤ 1, U(t) = e−iKψ

˜ AteiKψ At

can be derived from

slide-94
SLIDE 94

Conclusion/Outlook

general proofs of energy conditions in QFTs physical picture encoded in the entanglement structures (modular flow) holographic proof of QNEC using EWN: RT surface dynamics boundary modular flow “knows” about these… prescription for (fine-tuned classes of) modular flows in AdS/CFT

slide-95
SLIDE 95

Conclusion/Outlook

what happens in the “Milne wedges”? 1/N corrections to the prescription

  • ther bulk constraints from boundary modular flow, e.g. quantum

focusing conjecture (QFC)? Future directions:

Thank you!