Introduction to quantum computing

Assignment 1

Álvaro Maza Montesinos - Alejandro Vallejo Aparicio

April 25, 2025

1. Quantum theory

1.1

Let H2 denote the set of all complex Hermitian 2 × 2 matrices.

(a) Show that H2 is a real vector space.

(b) Show that {1, σx , σy , σz } (i.e. the identity and the Pauli matrices) is a basis of H2 as a real
vector space, i.e. A ∈ H2 if and only if it can be written uniquely as

A = r0 1 + rx σx + ry σy + rz σz ,

with r0 , rx , ry , rz ∈ R.
(c) Show that this basis is moreover orthogonal with respect to the (so-called Hilbert-Schmidt) inner
product
⟨A, B⟩ := tr(AT B).

(d) Prove that ρ is a qubit density matrix if and only if

1
ρ= 1 + rx σx + ry σy + rz σz
2
with r = (rx , ry , rz ) ∈ R3 such that ||r|| ≤ 1/2 (with equality if and only if ρ is a pure state).
Comment: This shows that the set of qubit density matrices can be represented by a ball in
R3 (with pure states corresponding to the sphere). This is known as the Bloch representation.
(e) Suppose a source emits n instances of the same unknown qubit state. Use the above to provide a
measurement scheme to reconstruct faithfully the density matrix produced by the source in the
limit n → ∞.

a) From the definition of H2 we know that ∀A ∈ H2 : A = A† . In order to show that H2 is a real

vector space we have to check that some conditions are satisfied.
But first, it is important to notice that the matrices in H2 are of the following type,
     
a z † a z a z
A= =A = = , ∀A ∈ H2 , where a, b ∈ R, z ∈ C (1)
z b z b z b

1
 Now that this is cleared out, we can go into check if the conditions are satisfied.

• Closed under the sum. This is, A + B → − C, where all the matrices are in H2 .
Knowing that A and B are. of course, in H2 , we can assure that they have the following
form, following eq.1,    
a z1 b z2
A= 1 B= 1 .
z1 a2 z2 b2
So, it is easy to check that,
     
a + b1 z1 + z2 a + b1 z1 + z2 c z3
A+B = 1 = 1 = 1 =C
z1 + z2 a2 + b2 z1 + z2 a2 + b2 z3 c2

Where C is clearly in H2 as it has the same form that we defined before for A and B.
• Closed under the multiplication by scalars. In this case we just have to check that if
A ∈ H2 , then c · A is still in H2 for c ∈ R. We directly check if c · A = (c · A)† .

c · A = (c · A)† = c · A† = c · A† = c · A.

Where we have used that c = c as c ∈ R.

• Commutativity A + B = B + A.
This can be easily check if we take a look at the previous result for A + B,
 
a + b1 z1 + z2
A+B = 1 ,
z1 + z2 a2 + b2

and from the commutativity of R and C, A + B = B + A.

It can also be checked in an easier way doing,

A + B = (A + B)† = A† + B † = B † + A† = B + A.

Where we have used the commutativity for the sum in general matrices.
• Associativity Again, using the definition that was given before (eq.1) and associativity
condition from R and C, it is easy to check that (A + B) + C = A + (B + C)
   
(a1 + b1 ) + c1 (z1 + z2 ) + z3 a1 + (b1 + c1 ) z1 + (z2 + z3 )
= .
(z1 + z2 ) + z3 (a2 + b2 ) + c2 z1 + (z2 + z3 ) a2 + (b2 + c2 )

• Existence of the null element This can be proved by finding the null element that in
this case is,  
0 0
0= .
0 0
0 + A = (0 + A)† = (A)† = A.

• Existence of the inverse element. Now it is sufficient to prove that if A ∈ H2 then

−A ∈ H2 .
−A = (−A)† = −A† = −A.

b) First of all, lets show which are the Pauli matrices,

     
0 1 0 −i 1 0
σx = , σy = , σz = . (2)
1 0 i 0 0 −1

From eq.1 we now the general form of the matrices in H2 , but we can also give this additional
definition expressed just in real parameters,
 
a x + iy
A= , where a, b, x, y ∈ R.
x − iy b

2
 Taking into account that the matrices as shown in this form depend only on 4 real parameters,
this coincides with r0 , rx , ry , rz . This is a good sign.
Now, we check the possible combinations,
         
1 0 0 1 0 −i 1 0 r + rz rx − iry
A = r0 + rx + ry + rz = 0 . (3)
0 1 1 0 i 0 0 −1 rx + iry r0 − rz

Which coincides with the definition we gave above.

The last thing to check is the linear independence of the different matrices that are involved in
the combination, for that we do
   
r0 + rz rx − iry 0 0
=
rx + iry r0 − rz 0 0

Solving this system, it is found that this is only possible if r0 = rx = ry = rz = 0. Thus we can
confirm that {1, σx , σy , σz } is a basis of H2 , then every matrix in H2 can be represented as a
unique linear combination of the elements of the basis, this is

A = r0 1 + rx σx + ry σy + rz σz .

c) Now, we are asked to show that the basis shown above is orthogonal with respect to the following
inner product,
⟨A, B⟩ = tr(A† B).
We just have to show that ⟨x, y⟩ = 0 for x, y whichever different elements of the basis. It becomes
easier after realising that as the space is H2 , the inner product can be expressed just as

⟨A, B⟩ = tr(A · B).

Now, we go onto showing that all the possible crossed products are zero.
First of all, because we are working with the Pauli matrices, it is clear that

⟨1, σi ⟩ = tr(1σi ) = tr(σi ) = 0, ∀i ∈ {x, y, z}.

Taking advantage of another property of the Pauli matrices ((σi )2 = 1), and the trace of 1 is
2. So, ⟨σi , σi ⟩ = 2 (the basis is not normal). Now checking the crossed inner products for the
different combinations, we find
 
i 0
⟨σx , σy ⟩ = tr(σx , σy ) = tr = 0.
0 −i
 
0 −1
⟨σx , σz ⟩ = tr(σx , σz ) = tr = 0.
1 0
 
0 i
⟨σy , σz ⟩ = tr(σy , σz ) = tr = 0.
i 0
Thus, the basis is orthogonal.
d) Here, we have to prove that ρ is a density matrix if and only if
1
ρ= 1 + rx σx + ry σy + rz σz .
2
We will start from
ρ = r0 1 + rx σx + ry σy + rz σz
and prove everything checking the different properties of a density matrix.
First, we know that for a density matrix, tr(ρ) = 1,

tr(ρ) = r0 · tr(1) + rx · tr(σx ) + ry · tr(σy ) + rz · tr(σz ) = 2r0 = 1

3
 then r0 = 12 .
The next condition that we will check is that ρ has to be positive semidefinite. This is that all
the eigenvalues of ρ have to be equal or greater than zero.
So, from eq.3 we calculate the eigenvalues. The calculus will not be included in this paper as it
is not of importance (simple eigenvalue problem), the result for the eigenvalues is,

λ = r0 ± (rx + ry + rz )1/2 .

Knowing that this has to be greater or equal to zero,

1
r0 − (rx + ry + rz )1/2 ≥ 0 →
− (rx + ry + rz )1/2 ≥ r0 = .
2
1
∥r∥ = (rx + ry + rz )1/2 ≥
.
2
Where we have chosen the minus (-) sign because the plus (+) is not of interest as it cannot give
a lower bound.
Now, the only thing left to do is prove that this inequality transforms into an equality when the
state is pure. The density matrix for a pure state satisfies ρ2 = ρ, so
1 1 1
ρ2 = ( 1 + ⃗r · ⃗σ )2 = 1 + ⃗r · ⃗σ + (⃗r · ⃗σ )2 = 1 + ⃗r · ⃗σ + ∥⃗r∥2 1.
2 4 4
1
ρ= 1 + ⃗r · ⃗σ .
2
1 1
ρ2 = ρ →
− 1 + ⃗r · ⃗σ + ∥⃗r∥2 1 = 1 + ⃗r · ⃗σ ,
4 2
1 1
∥⃗r∥2 1 = 1 → − ∥⃗r∥ = .
4 2
Where we have used that (⃗r · ⃗σ )2 = ∥⃗r∥2 .
e) From the past parts of this exercise we know that every ρ can be expressed as
1
ρ= 1 + rx σx + ry σy + rz σz .
2
In order to reconstruct the density matrix we just have to find ⃗r, as it has all the information.
For that purpose and taking into account that we have “infinite” amount of measurements, we
just have to divide the instances that we have into 3 groups (one for each Pauli matrix) and
measure ⟨σi ⟩. We know that

⟨σi ⟩ = tr(ρσi ) = tr(ri σi2 ) = 2ri .

⟨σi ⟩
So ri = 2 . With this information we can exactly reconstruct the density matrix.
To sum up, the measurement scheme would be:
• Divide the instances in three groups, in each of them we will make measurements for each
“direction”.
• Measure σi , the results are (±1)
Pn
• Find the expected value ⟨σi ⟩ = 1
n k=1 mk where mk = {±1}.
• In the limit n →
− ∞, we find exactly ⟨σi ⟩ = ri
2
• Doing this for each of the groups and we have found exactly ρ,

1 ⟨σx ⟩ ⟨σy ⟩ ⟨σz ⟩

ρ= 1+ σx + σy + σz .
2 2 2 2

4
1.2

Consider the 3-qubit state shared by A, B and C

1
|W ⟩ = √ (|001⟩ + |010⟩ + |100⟩).
3

(a) What is the reduced state of parties B and C? What is the reduced state for party A?
(b) Suppose A measures her qubit in the computational basis. What are the probabilities of each
possible outcome? What are the corresponding post-measurement states for ABC? And the
corresponding reduced post-measurement states for BC?
(c) Answer the same questions as above if A measures her qubit in the basis {|+⟩, |−⟩}.

a) We are given the 3-qubit state

1
|W ⟩ = √ (|001⟩ + |010⟩ + |100⟩),
3
where qubits are labelled as A, B, C, respectively.
To find the reduced state, we first find the reduced density matrix ρBC , we take the partial trace
over qubit A:
ρBC = trA (ρ)
1
ρ = |W ⟩⟨W | = (|001⟩ + |010⟩ + |100⟩)(⟨001| + ⟨010| + ⟨100|).
3
P1
Now, calculating the partial trace as ρBC = i=0 ⟨i|A ρ|i⟩A we find,

1
ρBC = (2|01⟩⟨01| + 2|10⟩⟨10| + |00⟩⟨00|).
3

This reduced density give us the following reduced state,

r
2 1
|ψ⟩BC = (|01⟩ + |10⟩) + √ |00⟩.
3 3
Now, following the same procedure for ρA = trBC (ρ), we find
1
ρA = (2|0⟩⟨0| + |1⟩⟨1|).
3
From this density matrix we obtain the reduced state,
r
2 1
|ψ⟩A = |0⟩ + √ |1⟩.
3 3

b) A measures the qubit in the computational basis {|0⟩ , |1⟩}

2 1
P (|0⟩A ) = , P (|1⟩A ) = ⇒ Observed directly from the coefficients in ρA (shown above).
3 3

This probabilities can also be obtained using the projective measure Pi = |i⟩⟨i|, then
2 1
P (|0⟩A ) = tr(|0⟩⟨0|ρA ) = , P (|1⟩A ) = tr(|1⟩⟨1|ρA ) = .
3 3

Now, we compute post-measurement state for ABC in the following manner:

5
 If the measured state is |0⟩A :

1
|W ′ ⟩ = √ (|001⟩ + |010⟩)
2
If the measured state is |1⟩A :

|W ′ ⟩ = |100⟩
For which we get that the reduced post-measured states are:
If we measure |0⟩A then

1
ρ′BC = (|01⟩ ⟨01| + |01⟩ ⟨10| + |10⟩ ⟨01| + |10⟩ ⟨10|)
2

1
⇒ ρ′BC = (|01⟩ + |10⟩) (⟨01| + ⟨10|)
2

1
⇒ |ψ ′ ⟩BC = √ (|01⟩ + |10⟩)
2
If we measure |1⟩A then

ρ′BC = |00⟩ ⟨00| , |ψ ′ ⟩BC = |00⟩

c) Let’s remember that
1 1
|+⟩ = √ (|0⟩ + |1⟩), |−⟩ = √ (|0⟩ − |1⟩).
2 2
Now we express |W ⟩ in the basis {|+⟩A , |−⟩A }:

1
|W ⟩ = √ (|0⟩ ⊗ (|01⟩ + |10⟩) + |1⟩ ⊗ |00⟩)
3
    
1 |+⟩ + |−⟩ |+⟩ − |−⟩
=√ √ ⊗ (|01⟩ + |10⟩) + √ ⊗ |00⟩
3 2 2
1
= √ (|+⟩ ⊗ (|10⟩ + |01⟩ + |00⟩) + |−⟩ ⊗ (|10⟩ + |01⟩ − |00⟩)) .
6

All the states have the same probability (can easily be seen just by looking at the coefficients).

1 1
P (|+⟩A ) = , P (|−⟩A ) = .
2 2
As before, this probabilities are also found using the projective measure,
1 1
P (|+⟩A ) = tr(|+⟩A ⟨+|A |W ⟩⟨W |) = , P (|−⟩A ) = tr(|−⟩A ⟨−|A |W ⟩⟨W |) = .
2 2
Now we compute the post-measurements:
If |+⟩A ⇒ |W ′ ⟩ = √1
3
|+⟩ ⊗ (|10⟩ + |01⟩ + |00⟩)
If |−⟩A ⇒ |W ′ ⟩ = √1
3
|−⟩ ⊗ (|10⟩ + |01⟩ − |00⟩)
and the reduced states are:

1
|W+′ ⟩BC = √ (|10⟩ + |01⟩ + |00⟩) .
3
1
|W−′ ⟩BC = √ (|10⟩ + |01⟩ − |00⟩) .
3

6
1.3

Let |ψ⟩ be an arbitrary pure n-qubit state, i.e.

X X
|ψ⟩ = αx1 ,...,xn |x1 · · · xn ⟩, |αx1 ,...,xn |2 = 1.
x1 ,...,xn =0,1 x1 ,...,xn =0,1

What is the probability of obtaining the outcome 0 on a measurement in the computational basis
on the first qubit? Conditioned on this outcome, what would be the post-measurement state of the
n-qubit system and the corresponding reduced state for the last n − 1 qubits?

The expression from the exercise can be expressed as

X X
|ψ⟩ = α0,x2 ,...,xn |0⟩ ⊗ |x2 , ..., xn ⟩ + α1,x2 ,...,xn |1⟩ ⊗ |x2 , ..., xn ⟩
x2 ,...,xn =0,1 x2 ,...,xn =0,1

From this expression it is easy to obtain the probability of measuring 0 on the first qubit,
X
P (0) = |α0,x2 ,...,xn |2 .
x2 ,...,xn =0,1

Conditioned to this outcome the post-measurement state would be

1 X
|ψ⟩ = P · α0,x2 ,...,xn |0⟩ ⊗ |x2 , ..., xn ⟩
[ x2 ,...,xn =0,1 |α0,x2 ,...,xn |2 ]1/2 x
2 ,...,xn =0,1

This can be expressed as the state for n − 1 qubits in the following way,
1 X
|ψ⟩ = P · α0,x2 ,...,xn |x2 , ..., xn ⟩
[ x2 ,...,xn =0,1 |α0,x2 ,...,xn |2 ]1/2 x
2 ,...,xn =0,1