An effect matrix with on finite vectors is a positive contraction on
ProvedBosonicLearn.effect_matrix_operatorA matrix whose finite quadratic forms lie between and is the matrix of a positive contraction on .
Let be any index set and let be a complex matrix such that for every finite set and every the quadratic form
is real and satisfies . Then there is a bounded operator on such that
- ;
- is positive;
- for every and every the series converges and equals ;
- for the standard basis vectors .
Role. The effects of a POVM on a Fock space are specified in this mission through their matrices and the bounds above. This statement turns each effect into an honest bounded operator, which is what is needed to pair effects with trace-class operators, to pass to limits, and to use operator-level results. It is a foundation for BosonicLearn.fockPOVM_postprocess and BosonicLearn.adaptive_scheme_product_povm.
Formalization Note The hypothesis is exactly the field effect_form of a FockPOVM at a measurable set. The space is Hilbert I, that is Mathlib's lp (fun _ : I => ℂ) 2; positivity is ContinuousLinearMap.IsPositive; clause 3 is a HasSum; is lp.single 2 i 1. The index set is arbitrary.
Related platform material The space used here (Hilbert A) is the same Lean type as OAI.TensorLp.H A, and is the same type as OAI.EntropyPhotonNumber.Fock n. The published definitions OAI_Analysis_TensorLp (tensor products, slices and operator lifts on ) and OAI_MathematicalPhysics_Fock_Operators (OAI.CrossEnsemble.operator, the bounded operator of a family, with hasSum_operator and entry_hasSum) are sorry-free and can be imported. A proved lemma of this mission that can be imported: BosonicLearn.traceDist_le_one (nuclear expansions, truncation to finite index sets).
import Mathlib import Definitions.Def_QAlgorithms_Defs_RealLabelRegisters open QAlgorithms.RealRegisters (Hilbert) open scoped ENNReal ComplexOrder
namespace BosonicLearn
/-- **An effect matrix is a bounded positive operator on `ℓ²`.**
Chen, Anna Mele, Anna Mele, Preskill, *When are bosonic Gaussian states classical to learn?*, arXiv:2609.26705, §1.1, p. 5: "All of our theorems stated in Sec. 1.2 hold for general POVM measurements on n-mode bosonic
systems". This makes precise the remark in the definition `BosonicLearn_FockPOVM` that an effect, given through
its quadratic forms on finitely supported vectors, "is a bounded operator between 0 and I".
Let `E = (E_{ij})` be a matrix indexed by a set `I` such that for every finite `s ⊆ I` and every `z` the form
`∑_{i,j ∈ s} conj(z_i) E_{ij} z_j` is real and lies between `0` and `∑_{i ∈ s} |z_i|²`. Then there is a bounded
operator `F` on `ℓ²(I)` with
1. `‖F‖ ≤ 1`;
2. `F` positive;
3. `(F z)_i = ∑_j E_{ij} z_j` for every `z ∈ ℓ²(I)` and every `i`, the series being convergent;
4. `⟨e_i, F e_j⟩ = E_{ij}` for the standard basis vectors.
**Formalization Note.** The hypothesis is exactly the field `effect_form` of a `FockPOVM` at a measurable set.
`ℓ²(I)` is `Hilbert I`, Mathlib's `lp (fun _ : I => ℂ) 2`; positivity is `ContinuousLinearMap.IsPositive`; clause 3
is a `HasSum`; `e_i` is `lp.single 2 i 1`. The index set is arbitrary. -/
theorem effect_matrix_operator {I : Type} [DecidableEq I] (E : Matrix I I ℂ)
(hE : ∀ (s : Finset I) (z : I → ℂ),
(∑ i ∈ s, ∑ j ∈ s, star (z i) * E i j * z j).im = 0 ∧
0 ≤ (∑ i ∈ s, ∑ j ∈ s, star (z i) * E i j * z j).re ∧
(∑ i ∈ s, ∑ j ∈ s, star (z i) * E i j * z j).re ≤ ∑ i ∈ s, ‖z i‖ ^ 2) :
∃ F : Hilbert I →L[ℂ] Hilbert I, ‖F‖ ≤ 1 ∧ F.IsPositive ∧
(∀ (z : Hilbert I) (i : I), HasSum (fun j => E i j * z j) (F z i)) ∧
∀ i j, inner ℂ (lp.single 2 i (1 : ℂ) : Hilbert I) (F (lp.single 2 j (1 : ℂ))) = E i j := by sorry
end BosonicLearn