Advanced Quantum

Information & Computation/SV HW 1: Warmup Homework PH601, January 8, 2025

x
   
1 0
• We represent the qubit states as |0⟩ = , |1⟩ =
0 1

1: Basic gates

Let’s recall from QIC-1 the basic gates of quantum computing for one, two and three qubits.

Single qubit gates

 
1 0
1. 1 = Identity gate = .
0 1
 
0 1
2. σx = X = = |0⟩⟨1| + |1⟩⟨0| |q⟩ X
1 0
 
0 −i
3. σy = Y = = −i(|0⟩⟨1| − |1⟩⟨0|) |q⟩ Y
i 0
 
1 0
4. σz = Z = = |0⟩⟨0| − |1⟩⟨1| |q⟩ Z
0 −1
 
1 1
5. Hadamard: H = √12 = √12 (|0⟩⟨0| + |0⟩⟨1| + |1⟩⟨0| − |1⟩⟨1|). |q⟩ H
1 −1

Two qubit gates

6. CNOT: UCNOT = |0⟩⟨0| ⊗ 1 + |1⟩⟨1| ⊗ X

|c⟩ |c⟩
≡
|q⟩ X |q⟩

7. Controlled-U gate: |0⟩⟨0| ⊗ 1 + |1⟩⟨1| ⊗ U

|c⟩

|ψ⟩ U

8. SWAP gate:
USWAP = |00⟩⟨00| + |01⟩⟨10| + |10⟩⟨01| + |11⟩⟨11|.

Three qubit gates

9. Toffoli or CCNOT gate:

(P00 + P01 + P10 ) ⊗ 1 + P11 ⊗ X; Pij ≡ |ij⟩⟨ij|

1
Advanced Quantum
Information & Computation/SV HW 1: Warmup Homework PH601, January 8, 2025

|c1 ⟩
|c2 ⟩

|q⟩

10. Fredkin or C-SWAP:

|0⟩⟨0| ⊗ 1 + |1⟩⟨1| ⊗ USWAP

|c⟩

|ψ1 ⟩

|ψ2 ⟩

2: CPTP Maps

Recall that a map Φ is CPTP iff:

1. Φ[ρ] ≥ 0
2. tr(Φ[ρ]) = tr(ρ)
3. Φ(ρ)† = Φ(ρ)

And if Φ ∈ CPTP, it has an OSR (Operator Sum Representation) given by:

X X
Φ(ρ) = Kµ ρKµ† s.t. Kµ† Kµ = 1.
µ µ

4: Choi isomorphism of a CPTP map

2
Recall once more the Choi isomorphism of a CPTP map: ∀ linear maps Φ : Hd → Hd , ∃ ρΦ ∈ H+d such that

1. ρΦ is isomorphic to Φ.
2. ρΦ ≥ 0, tr(ρΦ ) = 1 if Φ ∈ CPTP.

5: Stinespring Dilation

• For all Φ ∈ CPTP, we can write, for all ρ ∈ Hd :

 † 
Φ(ρS ) = trE USE (ρS ⊗ τE )USE . (1)
P
It is useful to note that if τE = k λk |λk ⟩⟨λk |, then:
X
Φ[ρS ] = λk Φk [ρS ]
k
† 
X 
= λk trE USE (ρS ⊗ |λk ⟩⟨λk |)USE . (2)
k

2
Advanced Quantum
Information & Computation/SV HW 1: Warmup Homework PH601, January 8, 2025

6: Naimark Extension

For every POVM {Ei } on a Hilbert space H, there exists a PVM {Pi } on a larger Hilbert space K such that the
POVM elements are obtained by taking the partial trace over the complementary subspace K′ of K, i.e.,

Ei = TrK′ (Pi ).

This extension is achieved through an isometry V : H → K with the property that Ei = TrK′ (V Ei V † ), where
V † V = IH , and the probabilities for measurement outcomes are given by Pi = V Ei V † .

{Note that we need Pauli basis and PTMS in the coming weeks. The rest of
the homework is just to jog your memory since some ideas such as Choi and
Stinespring are used in proofs.}

3
Advanced Quantum
Information & Computation/SV HW 1: Warmup Homework PH601, January 8, 2025

Solve the following questions (p̃ = (1 − p) and Pi ≡ |i⟩⟨i|):

1: States

1. Consider two qubit Wener states expressed in the Pauli basis,

1 
ρW = 14 − α(σ1x σ2x + σ1y σ2y + σ1z σ2z ) (3)
4
Express it in Gell-mann basis.
Gell-mann basis: The d– dimensional Gell-mann basis is given as
(k < l = 1, · · · d − 1),
(d)
Λ00 ≡ 1d , (4a)
r
(d) d
Λkl ≡ xkl = (|k⟩⟨l| + |l⟩⟨k|) , (4b)
r 2
(d) d
Λlk ≡ ykl = (−i|k⟩⟨l| + i|l⟩⟨k|) , (4c)
s 2 !
l−1
(d) d X
Λll ≡ zll = −l|l⟩⟨l| + |j⟩⟨j| . (4d)
l(l + 1) j=0

2. Repeat the same exercise for Bell-diagonal states expressed in Paui basis as,
1 
ρbd = 14 − t1 σ1x σ2x − t2 σ1y σ2y − t3 σ1z σ2z . (5)
4

2: Gates and channels

3. Find the matrix representation for Fredkin gate.

4. Find the Kraus operators for the following CPTP maps (system and
environment are highlighted):
(a) Bit-flip channel (p̃ = 1 − p):

4
Advanced Quantum
Information & Computation/SV HW 1: Warmup Homework PH601, January 8, 2025

(b) Phase flip channel

(c) Depolarising channel

5. Amplitude damping channel: The channel is given by the Kraus operators:

   
1 0 √ 0 1
E0 = √ & E1 = p (6)
0 p̃ 0 0
(a) Check it is CPTP.
(b) Dilate it into a circuit diagram.
Hint: Think what the role of the environment has to be and with
probability p, it has to spit out the P1 projector

3: Isomorphism

6. Isomorphism: Row and column stacking

P P
(a) Let ρ = ij ρij |i⟩⟨j| be mapped to |ρ⟩⟩ = ij ρij |ij⟩. Work out what
XρY is mapped to.
P
(b) Do the same for |ρ⟩⟩ = ij ρij |j⟩ ⊗ |i⟩.
7. t–qubit error channels: An error acting on t qubits is termed as a t-qubit
error. An n-qubit quantum channel is a t-qubit error channel if it has a
Kraus decomposition for which all Kraus operators are t-qubit errors.
Convince youself that the circuit shown in figure (1) is a 2– qubit error
channel.

4: Fixed points of CPTP maps

A fixed point ρf of a channel E is a state which is left unchanged by the action

of the channel so that:
E(ρf ) = ρf

5
Advanced Quantum
Information & Computation/SV HW 1: Warmup Homework PH601, January 8, 2025

Figure 1: Two-qubit error channel

8. (a) Consider H = ω0 σz . What is the fixed point of this CPTP map?

(b) Phase-flip map: Prove that the fixed points of a phase-flip map form a
convex set.
(c) Consider the Lindblad equation given by
dρ † 1 †
dt = D[X]ρ = XρX − 2 {X X, ρ}. here (D is the dissipator) and let
D[X]|0⟩ = 0. Prove that |0⟩ is a fixed point.
† −1
(X X)
(d) Evan’s theorem: If no such |0⟩ exists, prove tr{(X † X)−1 } is a fixed point.

(e) Unital maps: µ Kµ† Kµ = 1 and µ Kµ Kµ† = 1, then prove that

P P
1
Φ(1d ) = 1d (1d ≡ tr(1) ).

5: Different representations of CP maps

9. A and B matrices:
(a) The action of a CP can be given in terms of “A– matrix” through the
following equation:
A : ρi → ρf = A(ρi )
n
X
(ρf )r′ s′ = (Aρi )r′ s′ = Ar′ s′ ;rs (ρi )rs ,
r,s=1

where ρi and ρf are initial and final density matrices respectively.

If we define another realigned process matrix B as,
Br′ r;s′ s = Ar′ s′ ;rs .
(i) show that positivity of ρf requires B ≥ 0. (ii) Find A and B
matrices for bitflip and phase flip channels for a single qubit.

6
Advanced Quantum
Information & Computation/SV HW 1: Warmup Homework PH601, January 8, 2025

Suggested reading: For A and B matrices, you may refer to “Canonical

structure of A and B maps”
10. Pauli transfer matrix (PTM): The PTM representation of a quantum
operation E is a superoperator Λ with entries
1
Λij = Tr [Pi E(Pj )] ,
d
where Pi , Pj ∈ {1, X, Y, Z}.
(a) Find the Pauli transfer matrices Λbf and Λpf corresponding to a bit-flip
√ √
channel and phase flip channels with Kraus operators { p̃1, pX} and
√ √
{ p̃1, pZ} respectively.
(b) Find the Pauli transfer matrix corresponding to amplitude damping
channel Λad with Kraus operators given in equation (6).
(c) What is the difference between Λbf , Λpf and Λad ? What is its physical
significance?