Documentation

MIPRE.Background.LIDT.Co.Pasting.Bernoulli.FromHToG.MoveLemmas.Basic

Section 12 pasting: from-H-to-G move lemmas #

Tensor, positivity, and Cauchy--Schwarz helper lemmas for the adjacent paper chain: the counterpart of MIPRE/Background/LIDT/MIPStarRE/LDT/Pasting/Bernoulli/FromHToG/MoveLemmas/Basic.lean in the port of planning/c6b-plan.md (milestone M11, section "Port conventions").

The slice family is an IdxPolyFamily params 𝔓, so the completed slices, the half-products and the recurrence weights are local operators in 𝔓 (the vendored Op ΞΉ), and α΄΄ is star. The sandwich domination psd_contraction_comm_sandwich_le reads no state and holds in any C⋆-algebra with its order. The two state lemmas whose vendored state only evaluates, fromHToG_qSDDCore_symm (vendored ψ : QuantumState ΞΉ, applied to the joint state) and fromHToG_closenessOfIP_avgContext (vendored ψ : QuantumState (ΞΉ Γ— ΞΉ)), take a vector state V : VecState K with joint operators K β†’L[β„‚] K, to which callers pass strategy.state; the placement lemma fromHToG_ev_leftTensor_rightTensor_mono_right_of_nonneg_left takes the symmetric model S : SymModel 𝔓 K, in the vendored argument position. fromHToG_closenessOfIP_avgContext drops the unused normalization hypothesis _hψ : ψ.IsNormalized (section "Port conventions"), so its vendored callers' ψbi hnorm becomes strategy.state.

Not ported #

References #

In LionSR/MIPStarRE at commit 507e8122, not in this repository:

theorem MIPRE.LIDT.Co.Pasting.fromHToG_qSDDCore_symm {K : Type u_2} [NormedAddCommGroup K] [InnerProductSpace β„‚ K] [CompleteSpace K] {Outcome : Type u_3} [Fintype Outcome] (V : VecState K) (A B : Outcome β†’ K β†’L[β„‚] K) :
V.qSDDCore A B = V.qSDDCore B A

Symmetry of the raw pointwise state-dependent distance core. This local form is useful when orienting adjoint half-sandwich commutators for the second paper commutation step.

theorem MIPRE.LIDT.Co.Pasting.fromHToG_ev_leftTensor_rightTensor_mono_right_of_nonneg_left {𝔓 : Type u_1} [CStarAlgebra 𝔓] [PartialOrder 𝔓] [StarOrderedRing 𝔓] {K : Type u_2} [NormedAddCommGroup K] [InnerProductSpace β„‚ K] [CompleteSpace K] (S : SymModel 𝔓 K) {A B C : 𝔓} (hA : 0 ≀ A) (hBC : B ≀ C) :
S.ev (S.L A * S.R B) ≀ S.ev (S.L A * S.R C)

If A is PSD and B ≀ C, then the corresponding bipartite scalar expectations with left/right tensor placement are monotone in the right factor.

theorem MIPRE.LIDT.Co.Pasting.psd_contraction_comm_sandwich_le {𝔄 : Type u_3} [CStarAlgebra 𝔄] [PartialOrder 𝔄] [StarOrderedRing 𝔄] {S B : 𝔄} (hS0 : 0 ≀ S) (hS1 : S ≀ 1) (hB0 : 0 ≀ B) (hSB : Commute S B) :
S * B * S ≀ B

If S is a PSD contraction commuting with B, then S * B * S ≀ B. This formalizes the paper's eq:S-sandwich domination step without using explicit square roots.

theorem MIPRE.LIDT.Co.Pasting.fromHToGRecurrenceWeight_sandwich_base_le {𝔓 : Type u_1} [CStarAlgebra 𝔓] [PartialOrder 𝔓] [StarOrderedRing 𝔓] (params : MIPStarRE.LDT.Parameters) [MIPStarRE.LDT.FieldModel params.q] (family : IdxPolyFamily params 𝔓) (prefixLen : β„•) {tailLen : β„•} (Ο„tail : MIPStarRE.LDT.Pasting.GHatType tailLen) :
have S := fromHToGRecurrenceWeight params family prefixLen Ο„tail; S * family.averagedSubMeas.total * S ≀ family.averagedSubMeas.total

Paper eq:S-sandwich for the complete branch average G.

theorem MIPRE.LIDT.Co.Pasting.fromHToGRecurrenceWeight_sandwich_one_sub_base_le {𝔓 : Type u_1} [CStarAlgebra 𝔓] [PartialOrder 𝔓] [StarOrderedRing 𝔓] (params : MIPStarRE.LDT.Parameters) [MIPStarRE.LDT.FieldModel params.q] (family : IdxPolyFamily params 𝔓) (prefixLen : β„•) {tailLen : β„•} (Ο„tail : MIPStarRE.LDT.Pasting.GHatType tailLen) :
have S := fromHToGRecurrenceWeight params family prefixLen Ο„tail; S * (1 - family.averagedSubMeas.total) * S ≀ 1 - family.averagedSubMeas.total

Paper eq:S-sandwich for the incomplete branch average I - G.

theorem MIPRE.LIDT.Co.Pasting.fromHToG_gHatIdxMeas_outcome_isHermitian {𝔓 : Type u_1} [CStarAlgebra 𝔓] [PartialOrder 𝔓] [StarOrderedRing 𝔓] (params : MIPStarRE.LDT.Parameters) [MIPStarRE.LDT.FieldModel params.q] (family : IdxPolyFamily params 𝔓) (x : MIPStarRE.LDT.Fq params) (g : MIPStarRE.LDT.Pasting.GHatOutcome params) :
star ((gHatIdxMeas params family x).outcome g) = (gHatIdxMeas params family x).outcome g

Completed ĝ measurement outcomes are self-adjoint. This records the positivity-to-self-adjointness conversion used when orienting the adjoint half-sandwich commutator in the Mβ‚‚ β†’ M₃ move.

The reverse half-product is the adjoint of the ordered half-product.

theorem MIPRE.LIDT.Co.Pasting.fromHToG_gHatHalfProductOutcomeOperator_snoc {𝔓 : Type u_1} [CStarAlgebra 𝔓] [PartialOrder 𝔓] [StarOrderedRing 𝔓] (params : MIPStarRE.LDT.Parameters) [MIPStarRE.LDT.FieldModel params.q] (family : IdxPolyFamily params 𝔓) (n : β„•) (xs : MIPStarRE.LDT.PointTuple params n) (x : MIPStarRE.LDT.Fq params) (gs : MIPStarRE.LDT.Pasting.GHatTupleOutcome params n) (g : MIPStarRE.LDT.Pasting.GHatOutcome params) :
gHatHalfProductOutcomeOperator params family (n + 1) (Fin.snoc xs x) (Fin.snoc gs g) = gHatHalfProductOutcomeOperator params family n xs gs * (gHatIdxMeas params family x).outcome g

Ordered half-products satisfy a snoc recursion.

Reversing a tuple turns the ordered half-product into its adjoint.

theorem MIPRE.LIDT.Co.Pasting.fromHToG_closenessOfIP_avgContext {K : Type u_2} [NormedAddCommGroup K] [InnerProductSpace β„‚ K] [CompleteSpace K] {Question : Type u_3} {OutcomeA : Type u_4} {OutcomeB : Type u_5} [Fintype OutcomeA] [Fintype OutcomeB] (V : VecState K) (π’Ÿ : MIPStarRE.LDT.Distribution Question) (_hπ’Ÿ : βˆ‘ q ∈ π’Ÿ.support, π’Ÿ.weight q ≀ 1) (A B : Question β†’ OutcomeA β†’ K β†’L[β„‚] K) (C : Question β†’ OutcomeA β†’ OutcomeB β†’ K β†’L[β„‚] K) (Ξ³ : ℝ) (hAB : (MIPStarRE.LDT.avgOver π’Ÿ fun (q : Question) => V.qSDDCore (A q) (B q)) ≀ Ξ³) (hC : (MIPStarRE.LDT.avgOver π’Ÿ fun (q : Question) => βˆ‘ a : OutcomeA, V.ev ((βˆ‘ b : OutcomeB, C q a b) * star (βˆ‘ b : OutcomeB, C q a b))) ≀ 1) :
|(MIPStarRE.LDT.avgOver π’Ÿ fun (q : Question) => βˆ‘ a : OutcomeA, βˆ‘ b : OutcomeB, V.ev (C q a b * A q a)) - MIPStarRE.LDT.avgOver π’Ÿ fun (q : Question) => βˆ‘ a : OutcomeA, βˆ‘ b : OutcomeB, V.ev (C q a b * B q a)| ≀ √γ

Averaged-context variant of closenessOfIP: the contraction side condition is only required after averaging over the question distribution.