imaginary chaos
play

Imaginary chaos Janne Junnila (EPFL) Les Diablerets, February 10th - PowerPoint PPT Presentation

Imaginary chaos Janne Junnila (EPFL) Les Diablerets, February 10th 2020 joint work with Eero Saksman and Christian Webb; Juhan Aru, Guillaume Bavarez and Antoine Jego 1 Outline 1. Introduction 2. Moments 3. XOR-Ising model 4. Regularity,


  1. Imaginary chaos Janne Junnila (EPFL) Les Diablerets, February 10th 2020 joint work with Eero Saksman and Christian Webb; Juhan Aru, Guillaume Bavarez and Antoine Jego 1

  2. Outline 1. Introduction 2. Moments 3. XOR-Ising model 4. Regularity, densities and monofractality 2

  3. Introduction

  4. 𝑑 (ℝ 𝑒 ) we have π”½π‘Œ(πœ’)π‘Œ(πœ”) = ∫ πœ’(𝑦)πœ”(𝑧)𝐷(𝑦, 𝑧) 𝑒𝑦 𝑒𝑧. Log-correlated Gaussian fjelds Let us fjx a bounded simply connected domain 𝐸 βŠ‚ ℝ 𝑒 . Heuristic defjnition A Gaussian fjeld π‘ŒβˆΆ 𝐸 β†’ ℝ is called log-correlated if π”½π‘Œ(𝑦)π‘Œ(𝑧) = 𝐷(𝑦, 𝑧) ≔ log 1 where 𝑕 is regular. Caveat Such fjelds cannot be defjned pointwise and must instead be understood as distributions (generalized functions). This means that for all πœ’, πœ” ∈ 𝐷 ∞ 3 |𝑦 βˆ’ 𝑧| + 𝑕(𝑦, 𝑧)

  5. Log-correlated Gaussian fjelds Let us fjx a bounded simply connected domain 𝐸 βŠ‚ ℝ 𝑒 . Heuristic defjnition A Gaussian fjeld π‘ŒβˆΆ 𝐸 β†’ ℝ is called log-correlated if π”½π‘Œ(𝑦)π‘Œ(𝑧) = 𝐷(𝑦, 𝑧) ≔ log 1 where 𝑕 is regular. Caveat Such fjelds cannot be defjned pointwise and must instead be understood as distributions (generalized functions). This means that for all πœ’, πœ” ∈ 𝐷 ∞ 3 |𝑦 βˆ’ 𝑧| + 𝑕(𝑦, 𝑧) 𝑑 (ℝ 𝑒 ) we have π”½π‘Œ(πœ’)π‘Œ(πœ”) = ∫ πœ’(𝑦)πœ”(𝑧)𝐷(𝑦, 𝑧) 𝑒𝑦 𝑒𝑧.

  6. Log-correlated Gaussian fjelds β€’ We will always assume at least that β€’ 𝑕 ∈ 𝑀 1 (𝐸 Γ— 𝐸) ∩ 𝐷(𝐸 Γ— 𝐸) β€’ 𝑕 is bounded from above β€’ These properties are enough to ensure that π‘Œ exists. (Assuming that the kernel 𝐷 is positive defjnite.) 4

  7. Log-correlated Gaussian fjelds β€’ We will always assume at least that β€’ 𝑕 ∈ 𝑀 1 (𝐸 Γ— 𝐸) ∩ 𝐷(𝐸 Γ— 𝐸) β€’ 𝑕 is bounded from above β€’ These properties are enough to ensure that π‘Œ exists. (Assuming that the kernel 𝐷 is positive defjnite.) 4

  8. Example: Tie 2D GFF Defjnition The 0 -boundary GFF π›₯ in the domain 𝐸 is a Gaussian fjeld with the covariance 𝔽π›₯(𝑦)π›₯(𝑧) = 𝐻 𝐸 (𝑦, 𝑧) β€’ universality: β€’ appears in the scaling limit of various height function models, random matrices, QFT, … β€’ a recent characterisation: the only random fjeld with conformally invariant law and domain Markov property (+some moment condition) [BPR19] 5 where 𝐻 𝐸 is the Green’s function of the Dirichlet Laplacian in 𝐸 .

  9. Example: Tie 2D GFF Defjnition The 0 -boundary GFF π›₯ in the domain 𝐸 is a Gaussian fjeld with the covariance 𝔽π›₯(𝑦)π›₯(𝑧) = 𝐻 𝐸 (𝑦, 𝑧) β€’ universality: β€’ appears in the scaling limit of various height function models, random matrices, QFT, … β€’ a recent characterisation: the only random fjeld with conformally invariant law and domain Markov property (+some moment condition) [BPR19] 5 where 𝐻 𝐸 is the Green’s function of the Dirichlet Laplacian in 𝐸 .

  10. Example: Tie GFF in the unit square Figure 1: An approximation of the GFF in the unit square. 6

  11. fjelds π‘Œ π‘œ and normalize properly when taking the limit as π‘œ β†’ ∞ . For any given 𝛿 ∈ (0, √2𝑒) the functions 𝜈 π‘œ (𝑦) ≔ 𝑓 π›Ώπ‘Œ π‘œ (𝑦)βˆ’ 𝛿2 2 π”½π‘Œ π‘œ (𝑦) 2 converge to a random measure 𝜈 . We say that 𝜈 = 𝜈 𝛿 is a GMC measure Gaussian Multiplicative Chaos β€’ In various applications one is interested in measures formally of the β€’ To rigorously defjne them one has to approximate π‘Œ with regular Theorem/Defjnition ([Kah85; RV10; Sha16; Ber17]) associated to π‘Œ . 7 form 𝑓 π›Ώπ‘Œ(𝑦) 𝑒𝑦 where 𝛿 is a parameter.

  12. For any given 𝛿 ∈ (0, √2𝑒) the functions 𝜈 π‘œ (𝑦) ≔ 𝑓 π›Ώπ‘Œ π‘œ (𝑦)βˆ’ 𝛿2 2 π”½π‘Œ π‘œ (𝑦) 2 converge to a random measure 𝜈 . We say that 𝜈 = 𝜈 𝛿 is a GMC measure Gaussian Multiplicative Chaos β€’ In various applications one is interested in measures formally of the β€’ To rigorously defjne them one has to approximate π‘Œ with regular Theorem/Defjnition ([Kah85; RV10; Sha16; Ber17]) associated to π‘Œ . 7 form 𝑓 π›Ώπ‘Œ(𝑦) 𝑒𝑦 where 𝛿 is a parameter. fjelds π‘Œ π‘œ and normalize properly when taking the limit as π‘œ β†’ ∞ .

  13. Gaussian Multiplicative Chaos β€’ In various applications one is interested in measures formally of the β€’ To rigorously defjne them one has to approximate π‘Œ with regular Theorem/Defjnition ([Kah85; RV10; Sha16; Ber17]) associated to π‘Œ . 7 form 𝑓 π›Ώπ‘Œ(𝑦) 𝑒𝑦 where 𝛿 is a parameter. fjelds π‘Œ π‘œ and normalize properly when taking the limit as π‘œ β†’ ∞ . For any given 𝛿 ∈ (0, √2𝑒) the functions 𝜈 π‘œ (𝑦) ≔ 𝑓 π›Ώπ‘Œ π‘œ (𝑦)βˆ’ 𝛿2 2 π”½π‘Œ π‘œ (𝑦) 2 converge to a random measure 𝜈 . We say that 𝜈 = 𝜈 𝛿 is a GMC measure

  14. 8 ≲ ‖𝑔‖ 2 For any 𝑔 ∈ 𝐷 ∞ sequence is Cauchy in 𝑀 2 (𝛻) . β€’ Otherwise one can do a similar computation to show that the β€’ If 𝜈 π‘œ (𝑔) is a martingale this immediately shows convergence. 𝑒𝑦 𝑒𝑧 𝐸 2 1 A simple 𝑀 2 -computation Existence of GMC when 𝛿 ∈ (0, βˆšπ‘’) 𝑑 (𝐸) we have 𝐸 2 𝑔(𝑦)𝑔(𝑧)𝔽𝑓 π›Ώπ‘Œ π‘œ (𝑦)+π›Ώπ‘Œ π‘œ (𝑧)βˆ’ 𝛿2 2 π”½π‘Œ π‘œ (𝑦) 2 βˆ’ 𝛿2 2 π”½π‘Œ π‘œ (𝑧) 2 𝑒𝑦 𝑒𝑧 𝔽|𝜈 π‘œ (𝑔)| 2 = ∫ 𝐸 2 𝑔(𝑦)𝑔(𝑧)𝑓 𝛿 2 π”½π‘Œ π‘œ (𝑦)π‘Œ π‘œ (𝑧) 𝑒𝑦 𝑒𝑧 = ∫ 𝐸 2 𝑓 𝛿 2 log |π‘¦βˆ’π‘§| 𝑒𝑦 𝑒𝑧 = ∫ ∞ ∫ |𝑦 βˆ’ 𝑧| 𝛿 2 < ∞.

  15. 8 ≲ ‖𝑔‖ 2 For any 𝑔 ∈ 𝐷 ∞ sequence is Cauchy in 𝑀 2 (𝛻) . β€’ Otherwise one can do a similar computation to show that the β€’ If 𝜈 π‘œ (𝑔) is a martingale this immediately shows convergence. 𝑒𝑦 𝑒𝑧 𝐸 2 1 A simple 𝑀 2 -computation Existence of GMC when 𝛿 ∈ (0, βˆšπ‘’) 𝑑 (𝐸) we have 𝐸 2 𝑔(𝑦)𝑔(𝑧)𝔽𝑓 π›Ώπ‘Œ π‘œ (𝑦)+π›Ώπ‘Œ π‘œ (𝑧)βˆ’ 𝛿2 2 π”½π‘Œ π‘œ (𝑦) 2 βˆ’ 𝛿2 2 π”½π‘Œ π‘œ (𝑧) 2 𝑒𝑦 𝑒𝑧 𝔽|𝜈 π‘œ (𝑔)| 2 = ∫ 𝐸 2 𝑔(𝑦)𝑔(𝑧)𝑓 𝛿 2 π”½π‘Œ π‘œ (𝑦)π‘Œ π‘œ (𝑧) 𝑒𝑦 𝑒𝑧 = ∫ 𝐸 2 𝑓 𝛿 2 log |π‘¦βˆ’π‘§| 𝑒𝑦 𝑒𝑧 = ∫ ∞ ∫ |𝑦 βˆ’ 𝑧| 𝛿 2 < ∞.

  16. 8 ≲ ‖𝑔‖ 2 For any 𝑔 ∈ 𝐷 ∞ sequence is Cauchy in 𝑀 2 (𝛻) . β€’ Otherwise one can do a similar computation to show that the β€’ If 𝜈 π‘œ (𝑔) is a martingale this immediately shows convergence. 𝑒𝑦 𝑒𝑧 𝐸 2 1 A simple 𝑀 2 -computation Existence of GMC when 𝛿 ∈ (0, βˆšπ‘’) 𝑑 (𝐸) we have 𝐸 2 𝑔(𝑦)𝑔(𝑧)𝔽𝑓 π›Ώπ‘Œ π‘œ (𝑦)+π›Ώπ‘Œ π‘œ (𝑧)βˆ’ 𝛿2 2 π”½π‘Œ π‘œ (𝑦) 2 βˆ’ 𝛿2 2 π”½π‘Œ π‘œ (𝑧) 2 𝑒𝑦 𝑒𝑧 𝔽|𝜈 π‘œ (𝑔)| 2 = ∫ 𝐸 2 𝑔(𝑦)𝑔(𝑧)𝑓 𝛿 2 π”½π‘Œ π‘œ (𝑦)π‘Œ π‘œ (𝑧) 𝑒𝑦 𝑒𝑧 = ∫ 𝐸 2 𝑓 𝛿 2 log |π‘¦βˆ’π‘§| 𝑒𝑦 𝑒𝑧 = ∫ ∞ ∫ |𝑦 βˆ’ 𝑧| 𝛿 2 < ∞.

  17. Complex values of 𝛿 β„œ(𝛿) β„‘(𝛿) βˆšπ‘’ βˆ’βˆšπ‘’ βˆ’βˆš2𝑒 √2𝑒 Figure 2: The subcritical regime 𝐡 for 𝛿 in the complex plane. β€’ In fact, 𝛿 ↦ 𝜈 𝛿 (𝑔) is an analytic function on 𝐡 [JSW19]. β€’ The circle corresponds to the 𝑀 2 -phase – in particular it contains the whole subcritical part of the imaginary axis. 9

  18. Complex values of 𝛿 β„œ(𝛿) β„‘(𝛿) βˆšπ‘’ βˆ’βˆšπ‘’ βˆ’βˆš2𝑒 √2𝑒 Figure 2: The subcritical regime 𝐡 for 𝛿 in the complex plane. β€’ In fact, 𝛿 ↦ 𝜈 𝛿 (𝑔) is an analytic function on 𝐡 [JSW19]. β€’ The circle corresponds to the 𝑀 2 -phase – in particular it contains the whole subcritical part of the imaginary axis. 9

  19. Complex values of 𝛿 β„œ(𝛿) β„‘(𝛿) βˆšπ‘’ βˆ’βˆšπ‘’ βˆ’βˆš2𝑒 √2𝑒 Figure 2: The subcritical regime 𝐡 for 𝛿 in the complex plane. β€’ In fact, 𝛿 ↦ 𝜈 𝛿 (𝑔) is an analytic function on 𝐡 [JSW19]. β€’ The circle corresponds to the 𝑀 2 -phase – in particular it contains the whole subcritical part of the imaginary axis. 9

  20. Imaginary multiplicative chaos Theorem/Defjnition ([JSW18; LRV15]) converge in probability in 𝐼 βˆ’π‘’/2βˆ’πœ (ℝ 𝑒 ) to a random distribution 𝜈 . β€’ Applications: XOR-Ising model [JSW18], two-valued sets of the GFF [SSV19] and certain random fjelds constructed using the Brownian loop soup [CGPR19]. 10 Let 𝛾 ∈ (0, βˆšπ‘’) . Then the random functions 𝜈 π‘œ (𝑦) ≔ 𝑓 π‘—π›Ύπ‘Œ π‘œ (𝑦)+ 𝛾2 2 π”½π‘Œ π‘œ (𝑦) 2

  21. Imaginary multiplicative chaos Theorem/Defjnition ([JSW18; LRV15]) converge in probability in 𝐼 βˆ’π‘’/2βˆ’πœ (ℝ 𝑒 ) to a random distribution 𝜈 . β€’ Applications: XOR-Ising model [JSW18], two-valued sets of the GFF [SSV19] and certain random fjelds constructed using the Brownian loop soup [CGPR19]. 10 Let 𝛾 ∈ (0, βˆšπ‘’) . Then the random functions 𝜈 π‘œ (𝑦) ≔ 𝑓 π‘—π›Ύπ‘Œ π‘œ (𝑦)+ 𝛾2 2 π”½π‘Œ π‘œ (𝑦) 2

  22. Moments

Recommend


More recommend