Tensor networks and emergent geometry

Transcript

1. Tensor networks and emergent geometry
   Sivaramakrishnan Swaminathan
   Theory Group
   Department of Physics, The University of Texas at Austin
   11 August 2017
   Research presentation @ Vicarious AI
   

   View Slide

2. Based on
   B. Czech, P. H. Nguyen and S. Swaminathan
   A defect in holographic interpretations of tensor networks
   JHEP 1703, 090 (2017)
   doi:10.1007/JHEP03(2017)090 [arXiv:1612.05698 [hep-th]].
   

   View Slide

3. Quantum states and tensor networks
   

   View Slide

4. Primer on tensors
   

   View Slide

5. Quantum states ∼ probability distributions
   States are vectors in a Hilbert space, with ψ|ψ = 1
   Alternately, density matrices ρ ≡ |ψ ψ| with Tr ρ = 1
   Compute expectation values of operators
   O ≡ Tr [ρO] = ψ|O|ψ
   Entanglement ∼ Mutual Information
   Bell’s inequality
   

   View Slide

6. Many-body states are high-dimensional
   Joint distribution on N random variables =⇒ dim. ∼ exp (N)
   Would be nice to handle infinite systems
   This is why QM is hard, even though it’s linear
   

   View Slide

7. Exploit physical principles for efficiency?
   Goal: Model typical states
   eg: Lowest energy state; use “power method”
   Typical states form a tiny fraction of Hilbert space
   Locality =⇒ “area scaling” of entanglement
   Additional symmetries (translation, scale invariance)
   

   View Slide

8. Tensor networks
   Approximate joint distributions by a variational ansatz which allows
   efficient representation and computation
   Number of variational parameters scale favorably
   

   View Slide

9. Multiscale Entanglement Renormalization Ansatz
   

   View Slide

10. MERA: Constraints
    =
    u†
    u
    i j
    k l
    i j
    k l
    =
    w†
    w
    i
    j
    i
    j
    

    View Slide

11. MERA: Efficient computations
    Causal structure of influence simplifies computations
    

    View Slide

12. Computing the variational parameters
    Iteratively improve tensors layer-by-layer, by
    Ascending the Hamiltonian (forward-propagation)
    Descending the state (back-propagation)
    Alternating minimization to optimize tensors
    Note: Recent developments demonstrate faster learning techniques
    

    View Slide

13. Minimal Updates Proposal (MUP)
    Modeling scale invariant systems with a local defect
    Originally motivated by computational convenience
    We provide a principled justification (Boundary OPE)
    

    View Slide

14. Searching for principles
    MERA works fabulously well in practice!
    Renormalization group flow
    Connections to quantum gravity
    Link between MERA and deep learning
    

    View Slide

15. Quantum gravity
    

    View Slide

16. A hard problem
    Holy grail of fundamental physics for the past half-century
    If we naively combine gravity and quantum mechanics
    “Infinities” from marginalizing over infinitely many DOFs
    (dependence on prior; loss of predictivity)
    Physicists care about answers being finite and unique,
    so they may be compared with experiment.
    

    View Slide

17. Holographic quantum gravity
    Quantum Mechanics on "boundary" = Quantum Gravity in "bulk"
    (justifications from string theory)
    Figure from: https://commons.wikimedia.org/wiki/File:AdS3_(new).png
    

    View Slide

18. MERA ???
    ←→ Holography
    MERA models entanglement structure in quantum states
    Holographic spacetime maps entanglement structure
    Bulk geodesic length ∼
    = boundary entanglement
    (Ryu-Takayanagi formula)
    Emergent direction encodes scale-dependence of entanglement
    (Renormalization group flow)
    

    View Slide

19. Searching for a more direct relationship
    Lots of discussion over the last several years. . .
    I’ll summarize recent understanding, without detailing justifications
    MERA discretizes the integral transform of bulk geometry
    

    View Slide

20. Simplest example: hyperbolic space H2
    Full conformal symmetry
    Easy to verify correspondence with MERA
    (I’m happy to sketch the calculation if desired)
    

    View Slide

21. Our generalization: defect geometries
    Reduced symmetry: more nuanced duality; harder computations
    Proposed a novel generalization of the MUP: Rayed MERA
    

    View Slide

22. Summary
    Brief overview of tensor networks
    MERA networks
    Holographic quantum gravity
    Relation between tensor networks and holographic geometry
    

    View Slide

23. Thank you!
    

    View Slide

24. Other publications
    P. Agrawal, C. Kilic, S. Swaminathan and C. Trendafilova
    Secretly Asymmetric Dark Matter
    Phys. Rev. D 95, no. 1, 015031 (2017)
    doi:10.1103/PhysRevD.95.015031 [arXiv:1608.04745 [hep-ph]].
    C. Kilic and S. Swaminathan
    Can A Pseudo-Nambu-Goldstone Higgs Lead To
    Symmetry Non-Restoration?
    JHEP 1601, 002 (2016)
    doi:10.1007/JHEP01(2016)002 [arXiv:1508.05121 [hep-ph]]
    

    View Slide