AFTD Auto-Formalizing Theoretical Domains

is_channel_jamiolkowski

def

A matrix J on H_A tensor H_B is the Jamiołkowski matrix of a quantum channel if its partial transpose on the input space is positive semidefinite and has partial trace over the output space equal to the identity matrix on the input space.

Statement

def is_channel_jamiolkowski {dA dB : Type*} [Fintype dA] [DecidableEq dA] [Fintype dB] [DecidableEq dB]
    (J : Matrix (dA × dB) (dA × dB) ℂ) : Prop
Defined in
arXiv:2610.34972 Section 1.3, equation (7)
Verified
07 Nov 2026
Axioms
none listed
Built on 2

Read back from the Lean

Given finite types dA and dB (with decidable equality) and a matrix J : Matrix (dA × dB) (dA × dB) ℂ, is_channel_jamiolkowski J asserts that: 1. The matrix M : Matrix (dA × dB) (dA × dB) ℂ defined by M (a1, b1) (a2, b2) = J (a2, b1) (a1, b2) (the partial transpose of J with respect to dA) is positive semidefinite with respect to Complex.partialOrder (that is, M is Hermitian and for every vector x : dA × dB → ℂ, 0 ≤ star x ⬝ᵥ M ⬝ᵥ x under Complex.partialOrder). 2. The partial trace of M over the second system dB, Matrix.traceRight M, is equal to the identity matrix 1 on dA (i.e., Matrix.traceRight M = 1 : Matrix dA dA ℂ).

Written by a model that saw only the Lean, never the English above. If the two disagree, that is worth a challenge.

Lean source view module on GitHub

/-- A matrix J is the Jamiolkowski matrix of a quantum channel if its partial transpose is positive semidefinite and has trace over the output space equal to 1. -/
def is_channel_jamiolkowski {dA dB : Type*} [Fintype dA] [DecidableEq dA] [Fintype dB] [DecidableEq dB]
    (J : Matrix (dA × dB) (dA × dB) ℂ) : Prop := letI : PartialOrder ℂ := Complex.partialOrder
  (Matrix.of (fun (a1, b1) (a2, b2) => J (a2, b1) (a1, b2)) : Matrix (dA × dB) (dA × dB) ℂ).PosSemidef ∧
    Matrix.traceRight (Matrix.of (fun (a1, b1) (a2, b2) => J (a2, b1) (a1, b2))) = 1
See it in the graphDiscuss this resultChallenge it