Entries of POVM effects are fixed combinations of four Born probabilities:
ProvedBosonicLearn.effect_entry_bornEvery matrix entry of a POVM effect is a fixed complex combination of the outcome probabilities of four fixed states.
Let be an index set and . There are four density operators on and four complex numbers , depending only on and , such that and, for every POVM on with outcomes in any measurable space , effects and outcome laws , and for every measurable ,
For one can take the pure states with , , and ; for , and .
Role. Since the states and coefficients do not depend on the POVM, the identity holds simultaneously for a whole family of POVMs. If the outcome laws are Markov kernels, every entry is therefore measurable, and is measurable for measurable . Conversely the effects of a POVM are recovered from its outcome laws. Both uses occur in the proof of BosonicLearn.fockPOVM_compose.
Formalization Note is a FockPOVM I O and the probability is (P.law (τ k) B).toReal. The quantifier over the outcome type O : Type is inside the existential, so the same τ and coef serve all POVMs. This strengthens the second part of BosonicLearn.fockPOVM_form_measure, where the four measures depend on the POVM.
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. Proved lemmas of this mission that can be imported: BosonicLearn.traceDist_le_one (nuclear expansions, truncation to finite index sets) and BosonicLearn.effect_matrix_operator (effect matrices as bounded positive operators).
import Mathlib import Definitions.Def_BosonicLearn_FockPOVM open QAlgorithms.Papers.Independent.Dependency50.P33 (IsDensity) open MeasureTheory open scoped ENNReal Matrix
namespace BosonicLearn
/-- **Matrix entries of POVM effects are fixed combinations of four Born probabilities.**
Chen, Anna Mele, Anna Mele, Preskill, *When are bosonic Gaussian states classical to learn?*, arXiv:2609.26705, §1.1, p. 5 (general POVM measurements; outcome probabilities `Tr(E ρ)` in the Fock basis
`{|m⟩}` of p. 4).
For basis indices `a, b ∈ I` there are four density operators `τ₀, …, τ₃` on `ℓ²(I)` and four complex numbers
`c₀, …, c₃`, depending only on `a` and `b`, such that for **every** POVM `P` on `ℓ²(I)` (any outcome space) and
every measurable set `B` of outcomes
`E(B)_ab = ∑ₖ cₖ · Pr_P[outcome ∈ B | state τₖ]`,
and `∑ₖ cₖ = δ_ab`.
**Formalization Note.** The probabilities are `(P.law (τ k) B).toReal`. For `a ≠ b` the states are the pure
states `(|a⟩ + ω|b⟩)/√2`, `ω ∈ {1, −1, i, −i}` (polarization); for `a = b` they are `|a⟩⟨a|`. -/
theorem effect_entry_born {I : Type} [DecidableEq I] (a b : I) :
∃ (τ : Fin 4 → Matrix I I ℂ) (coef : Fin 4 → ℂ), (∀ c, IsDensity (τ c)) ∧
(∑ c, coef c = (1 : Matrix I I ℂ) a b) ∧
∀ (O : Type) [MeasurableSpace O] (P : FockPOVM I O) (B : Set O), MeasurableSet B →
P.effect B a b = ∑ c, coef c * (((P.law (τ c) B).toReal : ℝ) : ℂ) := by sorry
end BosonicLearn