File size: 5,676 Bytes
4a56cd6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
import OAI.Probability.FreeIsing.Brownian.ContinuousRowCompletion
import OAI.Probability.FreeIsing.Gibbs.CovarianceNativeRowCauchy
import OAI.Probability.FreeIsing.Ising.InnovationIndependence
import OAI.Probability.FreeIsing.Ising.MixedFiltering

namespace OAI

noncomputable section
open MeasureTheory ProbabilityTheory Filter
open scoped NNReal Topology BigOperators

namespace Problem367.IsingInnovation

variable {d : β„•} {A : Type*} [MeasurableSpace A]

theorem finiteCovariance_fullTime_interval_energy_cauchy
    (hd : 3 ≀ d) {beta : ℝ} {ΞΌ : Measure (TreeVertex d β†’ IsingSpin)}
    (hΞΌ : HasFreeIsingLaw (Real.tanh beta) ΞΌ) (hb : 0 ≀ beta)
    (hcrit : ((d - 1 : β„•) : ℝ) * Real.tanh beta ^ 2 ≀ 1)
    (Ξ½ : Measure A) {B : ℝβ‰₯0 β†’ A β†’ ℝ} (hB : IsPreBrownianReal B Ξ½)
    (hmB : βˆ€ t, Measurable (B t)) (hcB : βˆ€ a, Continuous (fun t => B t a))
    (v : TreeVertex d) {T : ℝ} (hT : 0 ≀ T) :
    let P := ΞΌ.prod (Measure.infinitePi (fun _ : TreeVertex d => Ξ½))
    (βˆ€ n, Summable (fun u => ∫ t in (0 : ℝ)..T,
      ∫ q, finiteCovariance beta B v u n t.toNNReal q ^ 2 βˆ‚P)) ∧
    (βˆ€ n k, Summable (fun u => ∫ t in (0 : ℝ)..T,
      ∫ q, (finiteCovariance beta B v u n t.toNNReal q -
        finiteCovariance beta B v u k t.toNNReal q) ^ 2 βˆ‚P)) ∧
    βˆ€ Ξ΅ > (0 : ℝ), βˆƒ N : β„•, βˆ€ n β‰₯ N, βˆ€ k β‰₯ N,
      (βˆ‘' u, ∫ t in (0 : ℝ)..T,
        ∫ q, (finiteCovariance beta B v u n t.toNNReal q -
          finiteCovariance beta B v u k t.toNNReal q) ^ 2 βˆ‚P) < Ξ΅ := by
  have h := hΞΌ.finiteGibbsCovariance_fullTime_interval_energy_cauchy
    hd hb hcrit (fun _ : TreeVertex d => Ξ½) (fun _ => hB)
    (fun _ => hmB) (fun _ => hcB) v hT
  simpa only [finiteCovariance, plantedArrayObservation,
    ArrayPosteriorStationarity.spinArray, add_comm] using h

theorem exists_continuous_covariance_integral
    (hd : 3 ≀ d) {beta : ℝ} {ΞΌ : Measure (TreeVertex d β†’ IsingSpin)}
    (hΞΌ : HasFreeIsingLaw (Real.tanh beta) ΞΌ) (hb : 0 ≀ beta)
    (hcrit : ((d - 1 : β„•) : ℝ) * Real.tanh beta ^ 2 ≀ 1)
    (Ξ½ : Measure A) {B : ℝβ‰₯0 β†’ A β†’ ℝ} (hB : IsPreBrownianReal B Ξ½)
    (hmB : βˆ€ t, Measurable (B t)) (hcB : βˆ€ a, Continuous (fun t => B t a))
    (v : TreeVertex d) :
    let P := ΞΌ.prod (Measure.infinitePi (fun _ : TreeVertex d => Ξ½))
    let F := ArrayPosteriorStationarity.filtration B hmB
    βˆƒ (I : β„• β†’ TreeVertex d β†’ ℝβ‰₯0 β†’ State d A β†’ ℝ)
      (Y : β„• β†’ ℝβ‰₯0 β†’ State d A β†’ ℝ) (Z : ℝβ‰₯0 β†’ State d A β†’ ℝ)
      (hI2 : βˆ€ n u t, MemLp (I n u t) 2 P)
      (hY2 : βˆ€ n t, MemLp (Y n t) 2 P) (_hZ2 : βˆ€ t, MemLp (Z t) 2 P),
      (βˆ€ n, IsStronglyPredictable F (Y n) ∧ Martingale (Y n) F P ∧
        (βˆ€α΅ q βˆ‚P, Continuous (fun t => Y n t q))) ∧
      (βˆ€ n u (N : β„•), 0 < N β†’ Tendsto (fun k => ∫ q,
        (BrownianGrid.uniformSum (innovation beta B u) (finiteCovariance beta B v u n)
          N k q - I n u N q) ^ 2 βˆ‚P) atTop (𝓝 0)) ∧
      (βˆ€ n t, HasSum (fun u => (hI2 n u t).toLp (I n u t))
        ((hY2 n t).toLp (Y n t))) ∧
      (βˆ€ n t, (∫ q, Y n t q ^ 2 βˆ‚P) =
        βˆ‘' u, ∫ s in (0 : ℝ)..(t : ℝ),
          ∫ q, finiteCovariance beta B v u n s.toNNReal q ^ 2 βˆ‚P) ∧
      (βˆ€ n m (N : β„•), 0 < N β†’ (∫ q, (Y n N q - Y m N q) ^ 2 βˆ‚P) =
        βˆ‘' u, ∫ s in (0 : ℝ)..(N : ℝ), ∫ q,
          (finiteCovariance beta B v u n s.toNNReal q -
            finiteCovariance beta B v u m s.toNNReal q) ^ 2 βˆ‚P) ∧
      IsStronglyPredictable F Z ∧ Martingale Z F P ∧
      (βˆ€α΅ q βˆ‚P, Continuous (fun t => Z t q)) ∧ Z 0 =ᡐ[P] 0 ∧
      (βˆ€ t, (∫ q, Z t q βˆ‚P) = 0) ∧
      (βˆ€ t, Tendsto (fun n => ∫ q, (Y n t q - Z t q) ^ 2 βˆ‚P) atTop (𝓝 0)) ∧
      (βˆ€ t : ℝβ‰₯0, Tendsto (fun n => βˆ‘' u, ∫ s in (0 : ℝ)..(t : ℝ),
        ∫ q, finiteCovariance beta B v u n s.toNNReal q ^ 2 βˆ‚P) atTop
          (𝓝 (∫ q, Z t q ^ 2 βˆ‚P))) ∧
      (βˆ€ T, Tendsto (fun n => ∫ q,
        MartingaleStrongMaximal.maximalSquare (Y n - Z) T q βˆ‚P) atTop (𝓝 0)) := by
  let : IsProbabilityMeasure ΞΌ := hΞΌ.isProbabilityMeasure
  let : IsProbabilityMeasure Ξ½ := hB.isGaussianProcess.isProbabilityMeasure
  let P := ΞΌ.prod (Measure.infinitePi (fun _ : TreeVertex d => Ξ½))
  have hW (u : TreeVertex d) := isBrownianReal_innovation hΞΌ Ξ½ hB hmB hcB u
  have hAd (u : TreeVertex d) :=
    (innovation_martingale hΞΌ Ξ½ hB hmB hcB u).stronglyAdapted
  have hcross (s t : ℝβ‰₯0) (hst : s ≀ t) : Pairwise fun u w =>
      IndepFun (fun q => innovation beta B u t q - innovation beta B u s q)
        (fun q => innovation beta B w t q - innovation beta B w s q) P := by
    intro u w huw
    exact (innovation_increment_law_and_indep hΞΌ Ξ½ hB hmB hcB
      (Y := fun _ => ()) s t hst measurable_const).2.1.indepFun huw
  have hrow (N : β„•) := finiteCovariance_fullTime_interval_energy_cauchy
    hd hΞΌ hb hcrit Ξ½ hB hmB hcB v (Nat.cast_nonneg N)
  exact BrownianArrayFiltration.exists_global_continuous_integral_limit_of_row_energy_cauchy
    (fun u => (hW u).toIsPreBrownianReal) hAd
    (innovation_increment_indep_filtration hΞΌ Ξ½ hB hmB hcB) hcross
    (fun u => (hW u).cont)
    (fun n u => finiteCovariance beta B v u n)
    (fun n u => finiteCovariance_stronglyAdapted beta B hmB v u n)
    (fun n u => Eventually.of_forall (finiteCovariance_continuous beta B hcB v u n))
    (fun _ _ => 2) (fun _ _ => by norm_num)
    (fun n u => Eventually.of_forall (fun q t => finiteCovariance_abs_le_two beta B v u n t q))
    (fun n N => (hrow N).1 n) (fun N _hN => (hrow N).2.2)

end Problem367.IsingInnovation

end

end OAI