Past exam of the mathematics course of the University of Cambridge / 2001 / iii / Paper 25 (Quantum Information Theory)

Use binary linear codes $C\subseteq C^{\prime }\subseteq {\mathbb{F}}_2^N$, with dimensions $k,k^{\prime }$. For each coset $u+C$ in $C^{\prime }/C$, defineRepresentatives of one coset give the same vector, and distinct cosets have disjoint computational supports, so these vectors are orthonormal. Their span is the CSS code $\mathcal{X}$, encoding $\boxed{k^{\prime }-k\text{ logical qubits in }N\text{ physical qubits}}$.

For $a,b\in {\mathbb{F}}_2^N$, write $X^a={⨂}_j{X}_j^{a_j}$ and $Z^b={⨂}_j{Z}_j^{b_j}$. The code is stabilized by $X^c$ for $c\in C$, because it permutes the sum, and by $Z^z$ for $z\in (C^{\prime }{)}^{\perp }$, because $z\cdot (u+c)=0$. These Pauli operators commute: $c\cdot z=0$. There are $k+N-k^{\prime }$ independent stabilizer generators, so their common positive eigenspace has dimension $2^{N-(k+N-k^{\prime })}=2^{k^{\prime }-k}$; it is exactly the span above.

The commuting check measurements separate the two types of error. A bit error $X^a$ changes the signs of the $Z$ checks by $(-1{)}^{z\cdot a}$, giving the ordinary syndrome of $a$ for $C^{\prime }$. A phase error $Z^b$ changes the $X$ checks by $(-1{)}^{c\cdot b}$, giving the syndrome of $b$ for the dual code $C^{\perp }$. If $C^{\prime }$ corrects $t_X$ bit errors and $C^{\perp }$ corrects $t_Z$ bit errors, these checks identify the respective patterns, and inverse Pauli operations recover the state without revealing the logical coset. This is the nested-code CSS syndrome recovery procedure.

The sharper distances account for harmless stabilizer errors. A Pauli $X^a{Z}^b$ commutes with every check precisely when $a\in C^{\prime }$ and $b\in C^{\perp }$. Among these, it acts as a stabilizer, up to a phase, precisely when $a\in C$ and $b\in (C^{\prime }{)}^{\perp }$. To justify the latter assertion, $X^a$ with $a\in C^{\prime }\setminus C$ permutes distinct logical cosets, while $Z^b$ with $b\in C^{\perp }\setminus (C^{\prime }{)}^{\perp }$ has unequal signs on two such cosets. Neither can act as a scalar, and a combined operator has the same nontrivial permutation or phase distinction. For positive logical dimension setThe quantum code distance is $\boxed{d=\min (d_X,d_Z)}$: a nontrivial logical Pauli has at least one nonstabilizer component of weight at least the corresponding distance, while a pure $X$ or pure $Z$ logical operator attains the minimum. The weight of $X^a{Z}^b$ counts the union of its supports, so this argument also includes $Y$ errors.

Here is the correction proof for coherent errors, not only a classical decoding argument. Let $P$ be the code projector and let $E,F$ have bit-component weights at most $t_X$ and phase-component weights at most $t_Z$. If $2t_X<d_X$ and $2t_Z<d_Z$, the corresponding components of $E^{†}F$ have weights below $d_X,d_Z$. If this product anticommutes with a stabilizer $S$, then $P(E^{†}F)P=-P(E^{†}F)P=0$, using $SP=P$. If it commutes with all stabilizers, the weight bounds force its two components into $C$ and $(C^{\prime }{)}^{\perp }$, so its code compression is a scalar multiple of $P$. Thus the Knill--Laflamme condition holds for the entire error family.

To see directly why this gives a common recovery, diagonalize its positive Gram matrix $c_{EF}$ and form the corresponding error combinations $F_j$. Their restrictions obey $P{F}_j^{†}{F}_{l}P=c_j{\delta }_{jl}P$. For $c_j>0$, $F_{j}P/\sqrt{c_j}$ is an isometry onto a syndrome subspace, and the different images are orthogonal. Measure these subspaces and apply the inverse isometry; zero-$c_j$ combinations annihilate the code. This recovers every quantum channel whose Kraus operators lie in the error span. Arbitrary operators supported on at most $t$ qubits expand in Pauli operators with the same support. Consequently the CSS code corrects all errors on up to $\boxed{t=\lfloor (d-1)/2\rfloor }$ qubits, and more generally the asymmetric family just described. The sufficient classical bound is obtained by replacing $d_X,d_Z$ by the minimum distances of $C^{\prime }$ and $C^{\perp }$, respectively; degeneracy can improve it.

The four Pauli matrices form an orthogonal basis of $2\times 2$ matrices for the Hilbert-Schmidt inner product, since $\mathrm{Tr}({\sigma }_{\mu }{\sigma }_{\nu })=2{\delta }_{\mu \nu }$. A Hermitian density matrix therefore has real coefficients in this basis, and its trace-one condition fixes the identity coefficient. ThusThe multiplication rule for the Pauli matrices ${\sigma }_i{\sigma }_j={\delta }_{ij}I+i{ϵ}_{ijk}{\sigma }_k$ implies $(n\cdot \sigma {)}^2=\Vert n{\Vert }^{2}I$. Its trace is zero, so for $n\ne 0$ its eigenvalues are $\pm \Vert n\Vert$; for $n=0$ it vanishes. HencePositivity is equivalent to $\Vert n\Vert \le 1$. Conversely every real vector in this unit ball defines a positive trace-one matrix by the boxed formula. Therefore $\Phi (\rho )=n$ is a one-to-one correspondence with the Bloch ball, with the displayed inverse. Each coordinate is linear in $\rho$, proving the affine property under any convex mixture.

A density matrix is pure precisely when its eigenvalues are $1,0$, equivalently when ${\rho }^2=\rho$. The eigenvalue formula makes this equivalent to $\Vert n\Vert =1$. Alternatively $\mathrm{Tr}{\rho }^2=(1+\Vert n{\Vert }^2)/2$. Thus the pure states correspond exactly to the Bloch sphere.

For $n\ne 0$, put $\hat{n}=n/\Vert n\Vert$. The spectral projectors for the larger and smaller eigenvalues areThey are orthogonal rank-one projections, have sum $I$, and give $\rho ={\lambda }_{+}{P}_{+}+{\lambda }_{-}{P}_{-}$. For normalized eigenvectors in this nondegenerate case they equal $|{\varphi }_1⟩⟨{\varphi }_1|$ and $|{\varphi }_2⟩⟨{\varphi }_2|$. Their Bloch vectors are therefore $\boxed{n_1=n/\Vert n\Vert ,n_2=-n/\Vert n\Vert }$, the antipodal eigenprojectors of a qubit density matrix.

At $n=0$, the state is $I/2$ and both eigenvalues coincide. There is no distinguished direction and the printed directional expression is undefined. Any orthonormal eigenbasis gives an antipodal pair of pure-state Bloch vectors, since its two projectors sum to $I$, but that pair can point in any direction. The nonzero-$n$ qualification is essential. The PDF denominators are $\Vert n\Vert$, not the converted TeX denominators involving $n_1,n_2$.

Use logarithms to base two and set $0{\log }_{2}0=0$. For a density matrix with eigenvalues $({\lambda }_j)$, the Von Neumann entropy isThus it is the Shannon entropy of the spectrum. In dimension $d$, each term is nonnegative, and entropy vanishes exactly when one eigenvalue is one, namely for a pure state. The Jensen inequality applied to the strictly concave function $-x{\log }_{2}x$ gives $S(\rho )\le {\log }_{2}d$, with equality precisely at the maximally mixed state $I/d$. Unitary conjugation preserves eigenvalues, so $S(U\rho U^{†})=S(\rho )$. Product eigenvalues giveThe finite-dimensional entropy is continuous because the eigenvalues depend continuously on the matrix and $-x\log x$ is continuous also at zero.

Define the partial trace over ${\mathcal{K}}_2$ by choosing an orthonormal basis $(e_b)$ and settingEquivalently it is uniquely characterized by $\mathrm{Tr}({\rho }_{1}A)=\mathrm{Tr}(\rho (A\otimes I))$ for every operator $A$ on ${\mathcal{K}}_1$. This identity proves basis independence. Define ${\rho }_2$ similarly, with $\mathrm{Tr}({\rho }_{2}B)=\mathrm{Tr}(\rho (I\otimes B))$. Hermiticity is preserved by the displayed sum. For each $u\in {\mathcal{K}}_1$,Taking the trace in a product orthonormal basis gives $\mathrm{Tr}{\rho }_1=\mathrm{Tr}\rho =1$, and the same proof works for ${\rho }_2$. Thus both reductions are density matrices. This is the partial trace positivity from product vectors argument.

For a bipartite pure state $\rho =|\psi ⟩⟨\psi |$, expand $\psi$ in a product orthonormal basis with coefficient matrix $C$. A singular value decomposition gives the Schmidt decompositionwith orthonormal families on both sides. Taking the two partial traces makes all cross terms vanish:The spectra may have different numbers of zeros but have identical nonzero eigenvalues. Consequently the reduced entropies coincide:This common value is the entanglement entropy. In infinite-dimensional trace-class systems the same partial-trace and Schmidt arguments apply, and the equality remains valid with the value $+\infty$ allowed; the bound ${\log }_{2}d$ is specifically finite-dimensional.

For completeness, the basic mixing and correlation properties also admit short proofs. Define quantum relative entropy by $D(\tau \Vert \sigma )=\mathrm{Tr}\tau ({\log }_2\tau -{\log }_2\sigma )$ when supports are included, and by $+\infty$ otherwise. Here is its nonnegativity proof without merely citing an operator inequality. Diagonalize $\tau ,\sigma$ with eigenvalues $r_i,s_j$, and let $q_{ij}$ be the squared overlaps of their eigenbases. Both row and column sums of $q$ are one. The scalar inequality $r\ln (r/s)\ge r-s$, with zero cases taken by limits, givesEquality forces $r_i=s_j$ on every nonzero overlap, hence $\tau =\sigma$. For $\overline{\rho }={\sum }_i{p}_i{\rho }_i$, expanding the trace definition yieldswhich proves Concavity of Von Neumann entropy.

The complementary entropy bounds for a quantum mixture areTo prove the upper bound, write the spectral pure-state decompositions of the ${\rho }_i$ as one ensemble with weights $w_{ij}=p_i{\lambda }_{ij}$. Form the column matrix with columns $\sqrt{w_{ij}}|{\psi }_{ij}⟩$. Its two Gram operators have the same nonzero spectrum: one is $\overline{\rho }$, and the other, $G$, has diagonal $w$. Relative entropy against $\mathrm{diag}w$ gives $0\le D(G\Vert \mathrm{diag}w)=H(w)-S(G)$. Since $H(w)=H(p)+{\sum }_i{p}_{i}S({\rho }_i)$ and $S(G)=S(\overline{\rho })$, the bound follows. Zero-weight columns are omitted.

Finally the support of $\rho$ is contained in $\mathrm{supp}{\rho }_1\otimes \mathrm{supp}{\rho }_2$. Indeed, a vector in a marginal kernel gives a sum of nonnegative product-vector expectations equal to zero; positivity forces the corresponding product vectors into the kernel of $\rho$. The operator identity $\log ({\rho }_1\otimes {\rho }_2)=\log {\rho }_1\otimes I+I\otimes \log {\rho }_2$ on this support and the partial-trace identities giveThis proves Subadditivity of Von Neumann entropy, with equality exactly for a product state. Purify ${\rho }_{12}$ by adjoining a third system, using its spectral decomposition. Equality of pure-state marginal entropies gives $S({\rho }_{13})=S({\rho }_2)$ and $S({\rho }_3)=S({\rho }_{12})$. Subadditivity on systems $1,3$ then gives $S({\rho }_2)\le S({\rho }_1)+S({\rho }_{12})$; swapping $1,2$ gives the Araki–Lieb inequality $\boxed{|S({\rho }_1)-S({\rho }_2)|\le S({\rho }_{12})}$. These finite-dimensional proofs establish the usual basic entropy bounds, rather than attributing all mixed-state behavior to pure-state Schmidt decompositions.

On an $N$-dimensional register, the quantum Fourier transform with positive-exponent convention isIt is the unitary version of the discrete Fourier transform on amplitudes. The inner product of columns $x,x^{\prime }$ is $N^{-1}{\sum }_y{e}^{2\pi i(x^{\prime }-x)y/N}$, equal to one when $x=x^{\prime }$ and zero otherwise by a finite geometric series, so the transformation is unitary.

For the qubit implementation take $N=2^n$, write $x=x_1\cdots x_n$ in most-significant-bit-first order, and put $0.x_j\cdots x_n={\sum }_{l=j}^n{x}_l{2}^{-(l-j+1)}$. Expanding $y={\sum }_j{y}_j{2}^{n-j}$ separates the Fourier phase as a product. The resulting dyadic quantum Fourier transform circuit identity isIndeed the phase on output bit $y_j$ is $e^{2\pi ix{y}_j/2^j}$, depending on the last $j$ input bits; multiplying the factors recovers $e^{2\pi ixy/2^n}$ and the normalization $2^{-n/2}$.

Implement the product in reverse wire order as follows. Process input wire $j$, starting at $j=1$. A Hadamard gate maps $|x_j⟩$ to $(|0⟩+e^{2\pi i{x}_j/2}|1⟩)/\sqrt{2}$. For each still-unprocessed wire $l>j$, apply a controlled phase gate with control $l$, target $j$, and target phaseOn a computational-basis input this multiplies the target's one-component by $e^{2\pi i{x}_l/2^{l-j+1}}$, so the processed wire becomes exactly $(|0⟩+e^{2\pi i0.x_j\cdots x_n}|1⟩)/\sqrt{2}$. Continue through all wires and reverse their order using SWAP gates. This equals the boxed transform on every basis vector and therefore on arbitrary superpositions by linearity; it does not require learning the input bits by measurement.

There are $\boxed{n\text{ Hadamards},n(n-1)/2\text{ controlled phase gates},\lfloor n/2\rfloor \text{ swaps}}$. A swap is three controlled-NOT gates, since the alternating control directions send $(a,b)$ successively to $(a,a\oplus b)$, then $(b,a\oplus b)$, and finally $(b,a)$. Thus the exact transform uses $O(n^2)$ one- and two-qubit gates from this phase-gate family, polynomial in the register size $n={\log }_{2}N$. This is a circuit-complexity statement, not a claim that all $N$ transformed amplitudes can be read out efficiently. If a fixed finite gate alphabet is required, the phase rotations must instead be synthesized to the desired accuracy; the exact count treats them as the available simple rotation gates.