Lean, informalized

Nielsen & Chuang · Chapter 12, Quantum information theory · p. 597

Exercise 12.34 - Projections onto a code space and its conjugate, a corrected error, and the norm of the result

Distilling Bell pairs by projecting onto a stabilizer code.

Hover over a blue or green term in the text, or a node in the graph, to find it in the other. Select it to read its declaration both in Lean and in words.

The Exercise

Alice and Bob share nn Bell pairs |β00⟩=(|00⟩+|11⟩)/2|\beta_{00}\rangle = (|00\rangle + |11\rangle)/\sqrt{2}; Alice holds one qubit of each pair and Bob the other. Without the normalization 2−n/22^{-n/2}, their joint state is |Φ⟩=∑x∈{0,1}n|x⟩A|x⟩B,|\Phi\rangle = \sum_{x \in \{0,1\}^n} |x\rangle_A \, |x\rangle_B ,(1)✦ the sum over the bit strings xx of length nn, with Alice’s kkth qubit paired with Bob’s kkth.

They distill entanglement with a stabilizer code on nn qubits. Alice projects her qubits onto the code space, and Bob projects his onto the complex conjugate of that space. Noise may first have struck Alice’s qubits with a Pauli error EE, which she undoes with a correction CC chosen from the measured syndrome before projecting. The exercise asks which vector the two projections leave, with and without the corrected error, and what its norm is.

The Pauli Group

The Pauli matrices are the four complex two by two matrices σ0=I\sigma_0 = I, σ1=X\sigma_1 = X, σ2=Y\sigma_2 = Y and σ3=Z\sigma_3 = Z, with rows and columns indexed by {0,1}\{0,1\} and X=(0110),Y=(0−ii0),Z=(100−1).X = \begin{pmatrix} 0 & 1 \\ 1 & 0 \end{pmatrix}, \qquad Y = \begin{pmatrix} 0 & -i \\ i & 0 \end{pmatrix}, \qquad Z = \begin{pmatrix} 1 & 0 \\ 0 & -1 \end{pmatrix}.(2)✦

For a string a=(a1,…,an)a = (a_1, \dots, a_n) of indices ak∈{0,1,2,3}a_k \in \{0,1,2,3\}, one for each of nn qubits, the Pauli string σa=σa1⊗⋯⊗σan\sigma_a = \sigma_{a_1} \otimes \cdots \otimes \sigma_{a_n} is the operator on the nn qubits whose matrix, with rows and columns labelled by bit strings xx and yy of length nn, has the entries (σa)xy=∏k=1n(σak)xkyk.(\sigma_a)_{xy} = \prod_{k=1}^{n} (\sigma_{a_k})_{x_k y_k} .(3)✦ The support of the string aa is the set of qubits kk with ak≠0a_k \neq 0.

The Pauli group GnG_n on nn qubits consists of the operators isσai^s \sigma_a: each element pp is given by its phase power ss, an integer modulo 44, and its Pauli string aa.✦ The phase isi^s is the imaginary unit raised to the representative of ss in {0,1,2,3}\{0,1,2,3\}, one of the four fourth roots of unity.✦ The matrix of pp is the complex 2n×2n2^n \times 2^n matrix isσai^s \sigma_a, and below an element stands for its matrix.✦ The element −I-I of GnG_n is the identity string with phase power 22.✦ The weight of an element pp of GnG_n is the number of qubits in the support of its string.✦

The Code

The stabilizer sum QQ of a subgroup SS of GnG_n is the sum ∑s∈Ss\sum_{s \in S} s of the matrices of its elements.✦ The code projector PSP_S of SS is the stabilizer sum divided by the number |S||S| of elements of SS, PS=Q|S|=1|S|∑s∈Ss,P_S = \frac{Q}{|S|} = \frac{1}{|S|} \sum_{s \in S} s ,(4)✦ the projector onto the code space VSV_S of the stabilizer code with stabilizer SS. The complex conjugate PS*P_S^{*} of the code projector PSP_S is the matrix whose entries are the complex conjugates of those of PSP_S; it is Bob’s projection.

The normalizer N(S)N(S) of SS is the set of elements gg of GnG_n with gSg−1=Sg S g^{-1} = S. The distance d(S)d(S) of the stabilizer code with stabilizer SS is the least weight of an element of N(S)N(S) that is not equal to irsi^r s for any element ss of SS and integer rr; when there is no such element, d(S)d(S) is infinite.✦

The Distilled State

The distilled vector |ψS⟩|\psi_S\rangle of a subgroup SS of GnG_n is the shared state (1) after Alice’s projection onto the code space: |ψS⟩=(PS⊗I)|Φ⟩.|\psi_S\rangle = (P_S \otimes I) \, |\Phi\rangle .(5)✦

The Claim

Theorem 1. Let SS be a subgroup of the Pauli group GnG_n.

(i) Projecting both halves gives the distilled vector: (PS⊗PS*)|Φ⟩=|ψS⟩.(P_S \otimes P_S^{*}) \, |\Phi\rangle = |\psi_S\rangle .(6)✦

(ii) Let tt be a natural number with d(S)≥2t+1d(S) \geq 2t + 1, and let EE and CC be elements of GnG_n of weight at most tt whose product CECE lies in N(S)N(S). Then there is a complex number cc with |c|=1|c| = 1 such that (PSCE⊗PS*)|Φ⟩=c|ψS⟩.(P_S \, C \, E \otimes P_S^{*}) \, |\Phi\rangle = c \, |\psi_S\rangle .(7)✦

(iii) Let m≤nm \leq n. If SS does not contain −I-I and has 2n−m2^{n-m} elements, then ⟨ψS|ψS⟩=2m.\langle \psi_S | \psi_S \rangle = 2^m .(8)✦

What Is Checked

The claim is stated over the distilled vector |ψS⟩|\psi_S\rangle of (5), built from the code projector PSP_S of (4), together with the matrices of the elements of GnG_n, their weights and the distance d(S)d(S). The distilled vector |ψS⟩|\psi_S\rangle is the unnormalized maximally entangled state on the code space VSV_S, that of mm encoded Bell pairs |β00⟩⊗m|\beta_{00}\rangle^{\otimes m} without the normalization.