Documentation

LeanPool.ParameterFreeGradient.V7.Proofs.Stage5AboveTwoLowerS5F.ObjectiveData

The normalized completion is a convex objective with the stated coordinate gradient.

noncomputable def V7.Stage5AboveTwoLowerS5F.unitDelta (p : ℝ) (T : ℕ) :

The normalized separation scale T ^ (-1 / p).

Equations
Instances For
    noncomputable def V7.Stage5AboveTwoLowerS5F.unitDeltaStep (p : ℝ) (T : ℕ) :

    The affine-piece offset used in the normalized resisting construction.

    Equations
    Instances For
      noncomputable def V7.Stage5AboveTwoLowerS5F.unitChi (p : ℝ) (T : ℕ) :

      The smoothing scale, equal to half the normalized affine-piece offset.

      Equations
      Instances For
        noncomputable def V7.Stage5AboveTwoLowerS5F.unitParameters (p : ℝ) (d T : ℕ) (algorithm : DeterministicExactPairAlgorithm d) (hT : 1 ≤ T) (hTd : T ≤ d) :

        The explicit kernel and scales initializing the normalized resisting construction.

        Equations
        • One or more equations did not get rendered due to their size.
        Instances For
          theorem V7.Stage5AboveTwoLowerS5F.unitDelta_pos {p : ℝ} {T : ℕ} (hT : 1 ≤ T) :
          0 < unitDelta p T
          theorem V7.Stage5AboveTwoLowerS5F.unitChi_pos {p : ℝ} {T : ℕ} (hT : 1 ≤ T) :
          0 < unitChi p T
          theorem V7.Stage5AboveTwoLowerS5F.unitBeta_pos {p : ℝ} {d T : ℕ} (hp : 2 < p) (hd : 2 ≤ d) (hT : 1 ≤ T) :
          noncomputable def V7.Stage5AboveTwoLowerS5F.unitCompletionData (p : ℝ) (d T : ℕ) (algorithm : DeterministicExactPairAlgorithm d) (hT : 1 ≤ T) (hTd : T ≤ d) :

          The complete normalized resisting data for a deterministic algorithm.

          Equations
          • One or more equations did not get rendered due to their size.
          Instances For
            theorem V7.Stage5AboveTwoLowerS5F.unitCompletionData_assumptions {p : ℝ} {d T : ℕ} (algorithm : DeterministicExactPairAlgorithm d) (hp : 2 < p) (hd : 2 ≤ d) (hT : 1 ≤ T) (hTd : T ≤ d) :
            theorem V7.Stage5AboveTwoLowerS5F.convexObjective_of_global_support {d : ℕ} (f : Point d → ℝ) (g : Point d → Point d) (hsupport : ∀ (x y : Point d), f x + pairing (g x) (y - x) ≤ f y) :

            A global supporting field implies convexity; this local wheel avoids any packaged convex-envelope dependency.

            theorem V7.Stage5AboveTwoLowerS5F.coordinateGradient_const_mul {d : ℕ} (f : Point d → ℝ) (g : Point d → Point d) (c : ℝ) (hgrad : O3.IsCoordinateGradient f g) :
            O3.IsCoordinateGradient (fun (x : O3.Vec d) => c * f x) fun (x : O3.Vec d) => c • g x
            noncomputable def V7.Stage5AboveTwoLowerS5F.unitObjectiveData (p : ℝ) (d T : ℕ) (algorithm : DeterministicExactPairAlgorithm d) (hT : 1 ≤ T) (hTd : T ≤ d) :

            The normalized completed resisting data viewed as objective data.

            Equations
            Instances For
              theorem V7.Stage5AboveTwoLowerS5F.unitObjectiveData_assumptions {p : ℝ} {d T : ℕ} (algorithm : DeterministicExactPairAlgorithm d) (hp : 2 < p) (hd : 2 ≤ d) (hT : 1 ≤ T) (hTd : T ≤ d) :