Documentation

MIPRE.Foundations.Introspection.PrefixLaw

The actual probability law of a CL prefix #

The weight is the pushforward of the uniform computational seed. Its Born and normalized-trace descriptions are proved, rather than supplied as extra premises. Restriction to the actual prefix image is a normalized distribution.

noncomputable def MIPRE.Introspection.readoutWeight {I : Type u_1} {X : Type u_2} [Fintype I] (f : I → X) (x : X) :

The law of a function of a uniform finite seed.

Equations
Instances For
    theorem MIPRE.Introspection.readoutWeight_eq_card {I : Type u_1} {X : Type u_2} [Fintype I] [DecidableEq I] [Fintype X] [DecidableEq X] (f : I → X) (x : X) :
    readoutWeight f x = ↑{i : I | f i = x}.card / ↑(Fintype.card I)
    theorem MIPRE.Introspection.readoutWeight_nonneg {I : Type u_1} {X : Type u_2} [Fintype I] [DecidableEq I] [Fintype X] [DecidableEq X] (f : I → X) (x : X) :
    theorem MIPRE.Introspection.sum_readoutWeight {I : Type u_1} {X : Type u_2} [Fintype I] [DecidableEq I] [Fintype X] [DecidableEq X] [Nonempty I] (f : I → X) :
    ∑ x : X, readoutWeight f x = 1
    theorem MIPRE.Introspection.readout_const_unit {I : Type u_1} [Fintype I] [DecidableEq I] :
    readout (fun (x : I) => ()) () = 1
    theorem MIPRE.Introspection.stateSqNorm_registerEPR_readout {I : Type u_1} {X : Type u_2} [Fintype I] [DecidableEq I] [Fintype X] [DecidableEq X] (f : I → X) (x : X) :

    The squared norm selected by a prefix projector is its actual uniform seed probability.

    def MIPRE.Introspection.prefixOutcomes {F : Type u_1} {ι : Type u_2} [Field F] [Fintype F] [DecidableEq F] [Fintype ι] [DecidableEq ι] {ℓ : ℕ} (P : CL.CLFun F ι ℓ) (k : ℕ) :
    Finset (ι → F)

    The actual possible outputs of the truncated CL sampler.

    Equations
    Instances For
      noncomputable def MIPRE.Introspection.prefixWeight {F : Type u_1} {ι : Type u_2} [Field F] [Fintype F] [Fintype ι] [DecidableEq ι] {ℓ : ℕ} (P : CL.CLFun F ι ℓ) (k : ℕ) (y : ι → F) :

      The paper's prefix distribution, before restricting to its support.

      Equations
      Instances For
        theorem MIPRE.Introspection.prefixWeight_eq_card {F : Type u_1} {ι : Type u_2} [Field F] [Fintype F] [DecidableEq F] [Fintype ι] [DecidableEq ι] {ℓ : ℕ} (P : CL.CLFun F ι ℓ) (k : ℕ) (y : ι → F) :
        prefixWeight P k y = ↑{z : ι → F | (CL.CLFun.truncate k P).eval z = y}.card / ↑(Fintype.card (ι → F))
        theorem MIPRE.Introspection.prefixWeight_nonneg {F : Type u_1} {ι : Type u_2} [Field F] [Fintype F] [DecidableEq F] [Fintype ι] [DecidableEq ι] {ℓ : ℕ} (P : CL.CLFun F ι ℓ) (k : ℕ) (y : ι → F) :
        theorem MIPRE.Introspection.sum_prefixWeight {F : Type u_1} {ι : Type u_2} [Field F] [Fintype F] [DecidableEq F] [Fintype ι] [DecidableEq ι] {ℓ : ℕ} (P : CL.CLFun F ι ℓ) (k : ℕ) :
        ∑ y : ι → F, prefixWeight P k y = 1
        theorem MIPRE.Introspection.prefixWeight_pos_iff {F : Type u_1} {ι : Type u_2} [Field F] [Fintype F] [DecidableEq F] [Fintype ι] [DecidableEq ι] {ℓ : ℕ} (P : CL.CLFun F ι ℓ) (k : ℕ) (y : ι → F) :
        theorem MIPRE.Introspection.prefixWeight_eq_zero {F : Type u_1} {ι : Type u_2} [Field F] [Fintype F] [DecidableEq F] [Fintype ι] [DecidableEq ι] {ℓ : ℕ} (P : CL.CLFun F ι ℓ) (k : ℕ) (y : ι → F) (hy : y ∉ prefixOutcomes P k) :
        prefixWeight P k y = 0
        theorem MIPRE.Introspection.sum_prefixWeight_supported {F : Type u_1} {ι : Type u_2} [Field F] [Fintype F] [DecidableEq F] [Fintype ι] [DecidableEq ι] {ℓ : ℕ} (P : CL.CLFun F ι ℓ) (k : ℕ) :
        ∑ y : ↥(prefixOutcomes P k), prefixWeight P k ↑y = 1

        Restricting the prefix law to attainable outputs preserves normalization.

        theorem MIPRE.Introspection.hidingPrefixOp_weight {F : Type u_1} {ι : Type u_2} [Field F] [Fintype F] [DecidableEq F] [Fintype ι] [DecidableEq ι] {ℓ : ℕ} [Algebra (ZMod 2) F] (P : CL.CLFun F ι ℓ) (k : ℕ) (y : ι → F) :

        The concrete honest prefix measurement has exactly the actual CL law.