Documentation

MIPRE.Background.LIDT.Co.MakingMeasurementsProjective.Orthonormalization.Completion

Option completion in the orthonormalization argument #

The auxiliary algebra used when the orthonormalization theorem is applied to the completion of a submeasurement by a fresh failure outcome, comparing the completed measurement with the restriction obtained by discarding this outcome: the counterpart of MIPRE/Background/LIDT/MIPStarRE/LDT/MakingMeasurementsProjective/Orthonormalization/Completion.lean in the port of planning/c6b-plan.md (milestone M8, section "Port conventions").

Both lemmas take a symmetric model S : SymModel 𝔓 K and a local submeasurement in 𝔓. The vendored hypotheses hperm : PermInvState ψ and hψ : ψ.IsNormalized of optionCompletion_bipartiteSSCRel are dropped: the swap step is S.ev_L_eq_ev_R and S.ev 1 = 1 is a theorem. The vendored A.liftLeft is A.map S.L. The vendored proof bounds the overlap of the totals through Preliminaries.qBipartiteSSCDefect_postprocess_le at the constant readout; here the same bound, βˆ‘β‚ ⟨Ψ, (Aₐ βŠ— Aₐ) Ψ⟩ ≀ ⟨Ψ, (T βŠ— T) Ψ⟩ for the total T, is the positivity of the off-diagonal overlaps, as in step 2 of SymModel.one_sub_sum_norm_sq_completion_le (Co/Doubling/Orthonormalization.lean).

Not ported #

Every declaration of the vendored file has a counterpart here.

Completing a submeasurement by a fresh failure outcome preserves bipartite strong self-consistency up to the paper's factor 2: the original diagonal gap controls the original outcomes, and the same gap controls the residual none outcome after applying the swap symmetry of the model.

Discarding the extra none outcome from the option-completed measurement can only decrease the qSDD sum: one simply drops a nonnegative summand.