OAI/Probability/DilutedSpin/ScheduledFrameError.lean

GitHub source file

7 declarations, 0 contain sorry, 0 declared as axioms

  1. theorem OAI.DilutedSpinGlass.ReducedTopology.shiftedFrameHeight line 14
  2. def OAI.DilutedSpinGlass.ReducedTopology.delayedFrame line 17

    Source documentation

    The literal two delayed target sides, retaining all old depth coordinates.

    def delayedFrame (H r d : ℕ) (k : ℕ+) (C : Fin k → ReducedTopology)
        (q : (j : Fin k) → (C j).Vertex → ℕ) : PrescribedTree (H+1+1+r+1+d) :=
    …

    Mentions: OAI.DilutedSpinGlass.ReducedTopology, OAI.DilutedSpinGlass.q, OAI.DilutedSpinGlass.ReducedTopology.Vertex, OAI.DilutedSpinGlass.PrescribedTree

  3. theorem OAI.DilutedSpinGlass.ReducedTopology.scheduled_frame_target_error_le line 26

    Source documentation

    Complete target projection comparison at the SAME height as the old frame. Both projectors and both entire target sides are the literal scheduled trees. No desired covariance estimate is an input.

    theorem scheduled_frame_target_error_le (H r d : ℕ) (k : ℕ+) (hk : 2≤(k:ℕ))
        (a : α) (C : Fin k → ReducedTopology)
        (e : (j : Fin k) → (C j).Vertex → {j : α // j≠a})
        (Q : α → Fin (H+1+1+r+1+d)) (ha : (Q a).val=r+1+d)
        (hQ : ∀ j v, (Q a).val+1≤(Q (e j v)).val)
        (T : KernelTower Ω (H+1+1+r+1+d))
        (f : FinitePath Ω (H+1+1+r+1+d) → Fin N → ℝ) (hf : ∀ x i, |f x i|≤1) :
        let ht := shiftedFrameHeight H r d
    …

    Mentions: OAI.DilutedSpinGlass.ReducedTopology, OAI.DilutedSpinGlass.ReducedTopology.Vertex, OAI.DilutedSpinGlass.KernelTower, OAI.DilutedSpinGlass.FinitePath, OAI.DilutedSpinGlass.ReducedTopology.shiftedFrameHeight

  4. lemma OAI.DilutedSpinGlass.kernelHeightCast_succ_fst line 95
  5. lemma OAI.DilutedSpinGlass.kernelHeightCast_succ_snd line 101
  6. lemma OAI.DilutedSpinGlass.vectorHeightCast_succ_pair line 108
  7. lemma OAI.DilutedSpinGlass.PrescribedTree.descendantEnergyAt_eq_shapeEnergyAt line 119

    Source documentation

    A proper-child covariance is the SAME single-copy covariance evaluated at the child's retained evaluation coordinate. Its two successive prefix averages are just one original-kernel prefix, not an independent pair of copies.

    lemma descendantEnergyAt_eq_shapeEnergyAt (C : PrescribedTree n) (r d : ℕ)
        (T : KernelTower Ω (n+1+r+1+d)) (f : FinitePath Ω (n+1+r+1+d) → Fin N → ℝ) :
        descendantEnergyAt C r d T f = shapeEnergyAt C (r+1+d)
          (kernelHeightCast (by omega) T) (vectorHeightCast (by omega) f) := by
    …

    Mentions: OAI.DilutedSpinGlass.PrescribedTree, OAI.DilutedSpinGlass.KernelTower, OAI.DilutedSpinGlass.FinitePath, OAI.DilutedSpinGlass.PrescribedTree.descendantEnergyAt, OAI.DilutedSpinGlass.PrescribedTree.shapeEnergyAt, OAI.DilutedSpinGlass.kernelHeightCast, OAI.DilutedSpinGlass.vectorHeightCast