Documentation

LeanPool.SingleCopyStabilizer.Round

A complete unsigned-label update for one binary round, including the zero-difference branch and a deterministic pivot choice.

@[reducible, inline]

Unsigned X or Z labels of a binary qubit register.

Equations
Instances For
    @[reducible, inline]

    Binary linear spaces of paired Pauli labels.

    Equations
    Instances For
      theorem LeanPool.SingleCopyStabilizer.exists_binary_pivot {ι : Type u_1} (u : BinaryVec ι) (hu : u ≠ 0) :
      ∃ (i : ι), u i = 1
      noncomputable def LeanPool.SingleCopyStabilizer.binaryPivot {ι : Type u_1} (u : BinaryVec ι) (hu : u ≠ 0) :
      ι

      Select a coordinate equal to one in a nonzero binary vector.

      Equations
      Instances For
        theorem LeanPool.SingleCopyStabilizer.binaryPivot_spec {ι : Type u_1} (u : BinaryVec ι) (hu : u ≠ 0) :
        u (binaryPivot u hu) = 1
        noncomputable def LeanPool.SingleCopyStabilizer.binaryRound {ι : Type u_1} [Fintype ι] [DecidableEq ι] (L : BinarySpace ι) (u : BinaryVec ι) (b : ZMod 2) :

        The observed shear and local correction acting on a binary label space.

        Equations
        Instances For
          @[simp]
          theorem LeanPool.SingleCopyStabilizer.binaryRound_of_ne {ι : Type u_1} [Fintype ι] [DecidableEq ι] (L : BinarySpace ι) (u : BinaryVec ι) (b : ZMod 2) (hu : u ≠ 0) :

          All computationally allowed differences preserve the diagonal dimension, including those which are not stabilizer X labels in the low-nullity setting.

          theorem LeanPool.SingleCopyStabilizer.binaryRound_exists_progress {ι : Type u_1} [Fintype ι] [DecidableEq ι] (L : BinarySpace ι) (hL : IsIsotropic L) (u : BinaryVec ι) (hu : u ∈ xProjection L) (hzero : u ≠ 0) :

          Every nonzero stabilizer X label has at least one successful correction.

          noncomputable def LeanPool.SingleCopyStabilizer.safeBinaryRound {ι : Type u_1} [Fintype ι] [DecidableEq ι] (L : BinarySpace ι) (u : BinaryVec ι) (b : ZMod 2) :

          Totalize sampling from the X projection by leaving the label space unchanged on labels outside that projection. This is used for exact learning; those labels can have positive probability for states of positive nullity.

          Equations
          Instances For