PH 510 / QU 511 Quantum computing assignment 1 solution

Due date: October 11th

1.

1) Check if 𝐻𝐻 is unitary.

Answer) 𝐻𝐻 † = |0⟩⟨+| + |1⟩⟨−|

⇒ 𝐻𝐻 † 𝐻𝐻 = |0⟩⟨0| + |1⟩⟨1| = 𝐼𝐼, 𝐻𝐻𝐻𝐻 † = |+⟩⟨+| + |−⟩⟨−| = 𝐼𝐼

2) Show explicitly that the Hadamard transform on 𝑛𝑛 qubits, 𝐻𝐻 ⊗𝑛𝑛 , may be written as
1
𝐻𝐻 ⊗𝑛𝑛 = � (−1)𝑥𝑥∙𝑦𝑦 |𝑥𝑥⟩⟨𝑦𝑦|,
√2𝑛𝑛 𝑥𝑥,𝑦𝑦∈{0,1}𝑛𝑛

Where {0,1}𝑛𝑛 represents the set of length-𝑛𝑛 binary strings.

1
Answer) For single qubit, in |0⟩, |1⟩ basis, 𝐻𝐻 = (|0⟩⟨0| + |1⟩⟨0| + |0⟩⟨1| − |1⟩⟨1|)
√2

1
= � (−1)𝑥𝑥∙𝑦𝑦 |𝑥𝑥⟩⟨𝑦𝑦|
√2 𝑥𝑥,𝑦𝑦∈{0,1}

∴ for n qubits,
1
𝐻𝐻 ⊗𝑛𝑛 = � (−1)𝑥𝑥∙𝑦𝑦 |𝑥𝑥⟩⟨𝑦𝑦|
√2𝑛𝑛 𝑥𝑥,𝑦𝑦∈{0,1}𝑛𝑛

2.

4 3
Answer) Let the matrix 𝐴𝐴 = � �.
3 4

To diagonalize 𝐴𝐴, det(𝐴𝐴 − 𝜆𝜆𝜆𝜆) = 0 ⇒ 𝜆𝜆 = 1,7

1 1
For 𝜆𝜆1 = 1, (𝐴𝐴 − 𝜆𝜆1 𝐼𝐼)|𝑣𝑣1 ⟩ = 0 ⇒ |𝑣𝑣1 ⟩ = � �
√2 −1
1 1
Similarly, for 𝜆𝜆2 = 7, (𝐴𝐴 − 𝜆𝜆2 𝐼𝐼)|𝑣𝑣2 ⟩ = 0 ⇒ |𝑣𝑣2 ⟩ = � �
√2 1
1 1 + √7 −1 + √7
∴ √𝐴𝐴 = �𝜆𝜆1 |𝑣𝑣1 ⟩⟨𝑣𝑣1 | + �𝜆𝜆2 |𝑣𝑣2 ⟩⟨𝑣𝑣2 | = � �,
2 −1 + √7 1 + √7
log 7 1 1
log 𝐴𝐴 = log 𝜆𝜆1 |𝑣𝑣1 ⟩⟨𝑣𝑣1 | + log 𝜆𝜆2 |𝑣𝑣2 ⟩⟨𝑣𝑣2 | = � �
2 1 1
3.

1) Show that an arbitrary single-qubit state 𝜌𝜌 can be written as

𝐼𝐼 + 𝑟𝑟⃗ ∙ 𝜎𝜎⃗
𝜌𝜌 = ,
2

Where 𝑟𝑟⃗ is a real three-dimensional vector such that ‖𝑟𝑟⃗‖ ≤ 1. This vector is called the
Bloch vector for the state 𝜌𝜌.
𝛼𝛼 𝛽𝛽
Answer) 𝜌𝜌 can be written as � ∗ � where 𝛼𝛼 + 𝛾𝛾 = 1.
𝛽𝛽 𝛾𝛾

Since 𝑟𝑟⃗ = �𝑟𝑟𝑥𝑥 , 𝑟𝑟𝑦𝑦 , 𝑟𝑟𝑧𝑧 �, 𝜎𝜎⃗ = �𝜎𝜎⃗𝑥𝑥 , 𝜎𝜎⃗𝑦𝑦 , 𝜎𝜎⃗𝑧𝑧 �,

𝐼𝐼 + 𝑟𝑟⃗ ∙ 𝜎𝜎⃗ 1 1 0 0 𝑟𝑟𝑥𝑥 0 −𝑖𝑖𝑖𝑖𝑦𝑦 𝑟𝑟 0 1 1 + 𝑟𝑟𝑧𝑧 𝑟𝑟𝑥𝑥 − 𝑖𝑖𝑟𝑟𝑦𝑦

= �� �+� �+� � + � 𝑧𝑧 �� = � �
2 2 0 1 𝑟𝑟
𝑥𝑥 0 𝑖𝑖𝑖𝑖𝑦𝑦 0 0 −𝑟𝑟𝑧𝑧 2 𝑟𝑟𝑥𝑥 + 𝑖𝑖𝑟𝑟𝑦𝑦 1 − 𝑟𝑟𝑧𝑧

We can compare these two matrices as 𝛼𝛼 = 1 + 𝑟𝑟𝑧𝑧 , 𝛽𝛽 = 𝑟𝑟𝑥𝑥 − 𝑖𝑖𝑟𝑟𝑦𝑦 , 𝛾𝛾 = 1 − 𝑟𝑟𝑧𝑧 .

1 + 𝑟𝑟𝑧𝑧 𝑟𝑟𝑥𝑥 − 𝑖𝑖𝑟𝑟𝑦𝑦 1 + 𝑟𝑟𝑧𝑧 𝑟𝑟𝑥𝑥 − 𝑖𝑖𝑟𝑟𝑦𝑦
� �� �
𝑟𝑟𝑥𝑥 + 𝑖𝑖𝑟𝑟𝑦𝑦 1 − 𝑟𝑟𝑧𝑧 𝑟𝑟𝑥𝑥 + 𝑖𝑖𝑟𝑟𝑦𝑦 1 − 𝑟𝑟𝑧𝑧

2) Write down the Bloch vector of the state 𝜌𝜌 = 𝐼𝐼 ⁄2.

Answer) We can easily check that 𝑟𝑟⃗ = (0,0,0).

3) Show that a state 𝜌𝜌 is pure if and only if ‖𝑟𝑟⃗‖ = 1.

Answer) State 𝜌𝜌 is pure if and only if 𝑇𝑇𝑇𝑇[𝜌𝜌2 ] = 1.

1 (1 + 𝑟𝑟𝑧𝑧 )2 + �𝑟𝑟𝑥𝑥2 + 𝑟𝑟𝑦𝑦2 � 2�𝑟𝑟𝑥𝑥 − 𝑖𝑖𝑟𝑟𝑦𝑦 � 1

𝜌𝜌2 = � 2 2 2
� ⇒ 𝑇𝑇𝑇𝑇[𝜌𝜌2 ] = �1 + 𝑟𝑟𝑥𝑥2 + 𝑟𝑟𝑦𝑦2 + 𝑟𝑟𝑧𝑧2 �
4 2�𝑟𝑟𝑥𝑥 + 𝑖𝑖𝑟𝑟𝑦𝑦 � (1 − 𝑟𝑟𝑧𝑧 ) + �𝑟𝑟𝑥𝑥 + 𝑟𝑟𝑦𝑦 � 2

∴ 𝑇𝑇𝑇𝑇[𝜌𝜌2 ] = 1 if and only if ‖𝑟𝑟⃗‖ = 1

4) For pure state case, we parametrize a qubit as

𝜃𝜃 𝜃𝜃
|𝜓𝜓⟩ = cos |0⟩ + 𝑒𝑒 𝑖𝑖𝑖𝑖 sin |1⟩
2 2
Write down the Bloch vector of the state.

Answer)
𝜃𝜃 𝜃𝜃 𝜃𝜃
cos 2 𝑒𝑒 −𝑖𝑖𝑖𝑖 sin cos
𝜌𝜌 = |𝜓𝜓⟩⟨𝜓𝜓| = � 2 2 2� =
𝜃𝜃 𝜃𝜃 𝜃𝜃
𝑒𝑒 𝑖𝑖𝑖𝑖 sin cos sin2
2 2 2
 1 1 + cos 𝜃𝜃 sin 𝜃𝜃 (cos 𝜙𝜙 − 𝑖𝑖 sin 𝜙𝜙)
� �
2 sin 𝜃𝜃 (𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 + 𝑖𝑖 sin 𝜙𝜙) 1 − cos 𝜃𝜃
∴ 𝑟𝑟⃗ = (𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 cos 𝜙𝜙 , sin 𝜃𝜃 sin 𝜙𝜙 , cos 𝜃𝜃)

4.

1) Prove that
𝜃𝜃 𝜃𝜃
𝑅𝑅𝑛𝑛� (𝜃𝜃) ≡ 𝑒𝑒 −𝑖𝑖𝑖𝑖𝑛𝑛� ∙𝜎𝜎�⃗⁄2 = cos 𝐼𝐼 − 𝑖𝑖 sin �𝑛𝑛𝑥𝑥 𝑋𝑋 + 𝑛𝑛𝑦𝑦 𝑌𝑌 + 𝑛𝑛𝑧𝑧 𝑍𝑍�
2 2
Where 𝑛𝑛� ∙ 𝜎𝜎⃗ = ∑3𝑖𝑖=1 𝑛𝑛𝑖𝑖 𝜎𝜎𝑖𝑖 and 𝜎𝜎𝑖𝑖 ’s are Pauli matrices.
𝑖𝑖𝑖𝑖
1 𝑖𝑖𝑖𝑖 𝑘𝑘
Answer) 𝑅𝑅𝑛𝑛�(𝜃𝜃) = 𝑒𝑒 − 2 �𝑛𝑛𝑥𝑥 𝑋𝑋+𝑛𝑛𝑦𝑦 𝑌𝑌+𝑛𝑛𝑧𝑧𝑍𝑍� = ∑∞
𝑘𝑘=0 𝑘𝑘! �− �𝑛𝑛𝑥𝑥 𝑋𝑋 + 𝑛𝑛𝑦𝑦 𝑌𝑌 + 𝑛𝑛𝑧𝑧 𝑍𝑍��
2

1 𝑖𝑖𝑖𝑖 𝑘𝑘
Since (𝑛𝑛� ∙ 𝜎𝜎⃗)2 = 𝐼𝐼, 𝑅𝑅𝑛𝑛�(𝜃𝜃) = ∑∞
𝑘𝑘=0 𝑘𝑘! �− �𝑛𝑛𝑥𝑥 𝑋𝑋 + 𝑛𝑛𝑦𝑦 𝑌𝑌 + 𝑛𝑛𝑧𝑧 𝑍𝑍��
2
∞ ∞
𝑙𝑙
1 𝜃𝜃 2𝑙𝑙 𝑚𝑚
1 𝜃𝜃 2𝑚𝑚+1
= �(−1) � 𝐼𝐼� − 𝑖𝑖 � (−1) � �𝑛𝑛 𝑋𝑋 + 𝑛𝑛𝑦𝑦 𝑌𝑌 + 𝑛𝑛𝑧𝑧 𝑍𝑍��
(2𝑙𝑙)! 2 (2𝑚𝑚 + 1)! 2 𝑥𝑥
𝑙𝑙=0 𝑚𝑚=0

𝜃𝜃 𝜃𝜃
= cos 𝐼𝐼 − 𝑖𝑖 sin �𝑛𝑛𝑥𝑥 𝑋𝑋 + 𝑛𝑛𝑦𝑦 𝑌𝑌 + 𝑛𝑛𝑧𝑧 𝑍𝑍�
2 2

2) Write down the matrix representations of the following operators

𝑅𝑅𝑥𝑥 (𝜃𝜃) ≡ 𝑒𝑒 −𝑖𝑖𝑖𝑖𝑖𝑖⁄2 , 𝑅𝑅𝑦𝑦 (𝜃𝜃) ≡ 𝑒𝑒 −𝑖𝑖𝑖𝑖𝑖𝑖⁄2 , 𝑅𝑅𝑧𝑧 (𝑡𝑡ℎ𝑒𝑒𝑒𝑒𝑒𝑒) ≡ 𝑒𝑒 −𝑖𝑖𝑖𝑖𝑖𝑖⁄2 .

𝜃𝜃 𝜃𝜃
𝜃𝜃 𝜃𝜃 cos −𝑖𝑖 sin
2 2
Answer) From 4-1), 𝑅𝑅𝑥𝑥 (𝜃𝜃) = cos 2 𝐼𝐼 − 𝑖𝑖 sin 𝑋𝑋 =� 𝜃𝜃 𝜃𝜃
�
2
−𝑖𝑖 sin cos
2 2

𝜃𝜃 𝜃𝜃 𝜃𝜃 𝜃𝜃
cos − sin cos − 𝑖𝑖 sin 0 𝑒𝑒 −𝑖𝑖𝑖𝑖 0�
Similarly, 𝑅𝑅𝑦𝑦 (𝜃𝜃) = � 𝜃𝜃2 𝜃𝜃
2
�, 𝑅𝑅𝑧𝑧 (𝜃𝜃) = � 2 2
𝜃𝜃 𝜃𝜃
� = �
sin cos 0 cos + 𝑖𝑖 sin 0 𝑒𝑒 𝑖𝑖𝑖𝑖
2 2 2 2

3) Show that the operator 𝑅𝑅𝑛𝑛� (𝜃𝜃) rotates the Bloch vector 𝑟𝑟⃗ by angle 𝜃𝜃 about the 𝑛𝑛�
axis of the Bloch sphere. This is why the operator is called rotation operator.

Answer) If we apply 𝑅𝑅𝑛𝑛� (𝜃𝜃) to a single qubit state 𝜌𝜌 with the Bloch vector 𝑟𝑟⃗, then

𝐼𝐼 + �𝑅𝑅𝑛𝑛� (𝜃𝜃)(𝑟𝑟⃗ ∙ 𝜎𝜎⃗)𝑅𝑅𝑛𝑛†� (𝜃𝜃)�

𝑅𝑅𝑛𝑛� (𝜃𝜃)𝜌𝜌𝑅𝑅𝑛𝑛†� (𝜃𝜃) = ,
2
�𝑅𝑅𝑛𝑛� (𝜃𝜃)(𝑟𝑟⃗ ∙ 𝜎𝜎⃗)𝑅𝑅𝑛𝑛†� (𝜃𝜃)�
 𝜃𝜃 𝜃𝜃 𝜃𝜃 𝜃𝜃
= �cos 𝐼𝐼 − 𝑖𝑖 sin (𝑛𝑛� ∙ 𝜎𝜎⃗)� (𝑟𝑟⃗ ∙ 𝜎𝜎⃗) �cos 𝐼𝐼 + 𝑖𝑖 sin (𝑛𝑛� ∙ 𝜎𝜎⃗)�
2 2 2 2

Since (𝑛𝑛� ∙ 𝜎𝜎⃗)(𝑟𝑟⃗ ∙ 𝜎𝜎⃗) = �𝑛𝑛𝑥𝑥 𝑋𝑋 + 𝑛𝑛𝑦𝑦 𝑌𝑌 + 𝑛𝑛𝑧𝑧 𝑍𝑍��𝑟𝑟𝑥𝑥 𝑋𝑋 + 𝑟𝑟𝑦𝑦 𝑌𝑌 + 𝑟𝑟𝑧𝑧 𝑍𝑍�

= �𝑛𝑛𝑥𝑥 𝑟𝑟𝑥𝑥 + 𝑛𝑛𝑦𝑦 𝑟𝑟𝑦𝑦 + 𝑛𝑛𝑧𝑧 𝑟𝑟𝑧𝑧 �𝐼𝐼 + 𝑖𝑖��𝑛𝑛𝑦𝑦 𝑟𝑟𝑧𝑧 − 𝑛𝑛𝑧𝑧 𝑟𝑟𝑦𝑦 �𝑋𝑋 + (𝑛𝑛𝑧𝑧 𝑟𝑟𝑥𝑥 − 𝑛𝑛𝑥𝑥 𝑟𝑟𝑧𝑧 )𝑌𝑌 + �𝑛𝑛𝑥𝑥 𝑟𝑟𝑦𝑦 − 𝑛𝑛𝑦𝑦 𝑟𝑟𝑥𝑥 �𝑍𝑍�

= (𝑛𝑛� ∙ 𝑟𝑟⃗)𝐼𝐼 + 𝑖𝑖(𝑛𝑛� × 𝑟𝑟⃗) ∙ 𝜎𝜎⃗,

𝜃𝜃 𝜃𝜃 𝜃𝜃 𝜃𝜃
�𝑅𝑅𝑛𝑛� (𝜃𝜃)(𝑟𝑟⃗ ∙ 𝜎𝜎⃗)𝑅𝑅𝑛𝑛†� (𝜃𝜃)� = �cos (𝑟𝑟⃗ ∙ 𝜎𝜎⃗) + sin {(𝑛𝑛� × 𝑟𝑟⃗) ∙ 𝜎𝜎⃗ − 𝑖𝑖(𝑛𝑛� ∙ 𝑟𝑟⃗)𝐼𝐼}� �cos 𝐼𝐼 + 𝑖𝑖 sin (𝑛𝑛� ∙ 𝜎𝜎⃗)�
2 2 2 2
𝜃𝜃 𝜃𝜃 𝜃𝜃 𝜃𝜃
= cos2 (𝑟𝑟⃗ ∙ 𝜎𝜎⃗) + 2 sin cos (𝑛𝑛� × 𝑟𝑟⃗) ∙ 𝜎𝜎⃗ + sin2 {(𝑟𝑟⃗ ∙ 𝑛𝑛�)(𝑛𝑛� ∙ 𝜎𝜎⃗) + 𝑛𝑛� × (𝑛𝑛� × 𝑟𝑟⃗) ∙ 𝜎𝜎⃗}
2 2 2 2
𝜃𝜃 𝜃𝜃 𝜃𝜃 𝜃𝜃
= �cos 2 𝑟𝑟⃗ + 2 sin cos (𝑛𝑛� × 𝑟𝑟⃗) + sin2 {(𝑟𝑟⃗ ∙ 𝑛𝑛�)𝑛𝑛� + 𝑛𝑛� × (𝑛𝑛� × 𝑟𝑟⃗)}� ∙ 𝜎𝜎⃗ ≡ 𝑅𝑅�⃗ ∙ 𝜎𝜎⃗
2 2 2 2
�⃗ ∙𝜎𝜎
𝐼𝐼+𝑅𝑅 �⃗
Thus, 𝑅𝑅𝑛𝑛� (𝜃𝜃)𝜌𝜌𝑅𝑅𝑛𝑛†� (𝜃𝜃) = 2
⇒ transformed single qubit state.
1+cos 𝜃𝜃 1−cos 𝜃𝜃
Also, 𝑅𝑅�⃗ = 𝑟𝑟⃗ + sin 𝜃𝜃 (𝑛𝑛� × 𝑟𝑟⃗) + {(𝑟𝑟⃗ ∙ 𝑛𝑛�)𝑛𝑛� + (𝑛𝑛� ∙ 𝑟𝑟⃗)𝑛𝑛� − (𝑛𝑛� ∙ 𝑛𝑛�)𝑟𝑟⃗}
2 2

= 𝑟𝑟⃗ cos 𝜃𝜃 + (𝑛𝑛� × 𝑟𝑟⃗) sin 𝜃𝜃 + (1 − cos 𝜃𝜃)(𝑛𝑛� ∙ 𝑟𝑟⃗)𝑛𝑛� which is a rotated vector 𝑟𝑟⃗ by angle 𝜃𝜃 about
the 𝑛𝑛� axis.

4) Express the Hadamard gate 𝐻𝐻 as a product of rotation operators.

Answer) Any expression that can generate the Hadamard gate would be the answer.
𝜋𝜋 1 0 −𝑖𝑖 1 −1 1 −𝑖𝑖 −𝑖𝑖 1 1 1
Example) 𝑅𝑅𝑥𝑥 (𝜋𝜋)𝑅𝑅𝑦𝑦 � � = � �� � = 2� � = −𝑖𝑖 × 2 � �
2 √2 −𝑖𝑖 0 1 1 √ −𝑖𝑖 𝑖𝑖 √ 1 −1

5.

1) Find a simple expression of �𝜎𝜎𝑖𝑖 , 𝜎𝜎𝑗𝑗 � and �𝜎𝜎𝑖𝑖 , 𝜎𝜎𝑗𝑗 � for 𝑖𝑖, 𝑗𝑗 = 0,1,2,3. Here, [∙,∙] and {∙,∙}
denote commutator and anticommutator.
𝐼𝐼, 𝑖𝑖 = 𝑗𝑗
Answer) 𝜎𝜎𝑖𝑖 𝜎𝜎𝑗𝑗 = �𝜀𝜀 × 𝑖𝑖𝜎𝜎 , 𝑖𝑖 ≠ 𝑗𝑗 ⇒ �𝜎𝜎𝑖𝑖 , 𝜎𝜎𝑗𝑗 � = 2𝑖𝑖𝜀𝜀𝑖𝑖𝑖𝑖𝑖𝑖 × 𝑖𝑖𝜎𝜎𝑘𝑘 , �𝜎𝜎𝑖𝑖 , 𝜎𝜎𝑗𝑗 � = 2𝛿𝛿𝑖𝑖𝑖𝑖 𝜎𝜎|𝑖𝑖−𝑗𝑗| .
𝑖𝑖𝑖𝑖𝑖𝑖 𝑘𝑘

Here, 𝜀𝜀𝑖𝑖𝑖𝑖𝑖𝑖 is a Levi-Civita symbol and 𝛿𝛿𝑖𝑖𝑖𝑖 is a Kronecker delta.

2) Prove that Pauli matrices form an orthogonal basis of ℂ2×2 with respect to the
Hilbert-Schmidt norm 〈𝐴𝐴, 𝐵𝐵〉𝐻𝐻𝐻𝐻 = 𝑇𝑇𝑇𝑇�𝐴𝐴† 𝐵𝐵�.
2, 𝑖𝑖 = 𝑗𝑗
Answer) Since 𝜎𝜎𝑖𝑖† = 𝜎𝜎𝑖𝑖 , 𝑇𝑇𝑇𝑇�𝜎𝜎𝑖𝑖† 𝜎𝜎𝑗𝑗 � = 𝑇𝑇𝑇𝑇�𝜎𝜎𝑖𝑖 𝜎𝜎𝑗𝑗 � = �
0, 𝑖𝑖 ≠ 𝑗𝑗
∴ Pauli matrices are orthogonal under Hilbert-Schmidt inner product.
1
⇒We can construct any matrix 𝑀𝑀 ∈ ℂ2×2 as 𝑀𝑀 = ∑3𝑖𝑖=0 𝑎𝑎𝑖𝑖 𝜎𝜎𝑖𝑖 with 𝑎𝑎𝑖𝑖 = 𝑇𝑇𝑇𝑇[𝑀𝑀𝜎𝜎𝑖𝑖 ].
2

3) Prove that for any 2 × 2 matrix 𝑀𝑀,

3

� 𝜎𝜎𝑖𝑖 𝑀𝑀𝜎𝜎𝑖𝑖 ∝ 𝜎𝜎0 .

𝑖𝑖=0

And decide the normalization factor to make the proportionality become equality.

Answer) ∑3𝑖𝑖=0 𝜎𝜎𝑖𝑖 𝑀𝑀𝜎𝜎𝑖𝑖 = ∑3𝑖𝑖=0 𝜎𝜎𝑖𝑖 �∑3𝑗𝑗=0 𝑎𝑎𝑗𝑗 𝜎𝜎𝑗𝑗 �𝜎𝜎𝑖𝑖 = 4𝑎𝑎0 𝜎𝜎0 = 2𝑇𝑇𝑇𝑇[𝑀𝑀] ∝ 𝜎𝜎0

𝑛𝑛 ×2𝑛𝑛
4) Prove that the tensor product of Pauli matrices form an orthogonal basis of ℂ2
with respect to the Hilbert-Schmidt norm.

Answer) 〈𝜎𝜎𝑖𝑖1 ⊗ 𝜎𝜎𝑖𝑖2 ⊗ … ⊗ 𝜎𝜎𝑖𝑖𝑛𝑛 , 𝜎𝜎𝑗𝑗1 ⊗ 𝜎𝜎𝑗𝑗2 ⊗ … ⊗ 𝜎𝜎𝑗𝑗𝑛𝑛 〉 = 2𝑛𝑛 𝛿𝛿𝑖𝑖1 𝑗𝑗1 𝛿𝛿𝑖𝑖2 𝑗𝑗2 … 𝛿𝛿𝑖𝑖𝑛𝑛 𝑗𝑗𝑛𝑛

5) Prove that for any 2𝑛𝑛 × 2𝑛𝑛 matrix 𝑀𝑀,

3 3 3

� � … � �𝜎𝜎𝑖𝑖1 ⊗ 𝜎𝜎𝑖𝑖2 ⊗ … ⊗ 𝜎𝜎𝑖𝑖𝑛𝑛 �𝑀𝑀�𝜎𝜎𝑖𝑖1 ⊗ 𝜎𝜎𝑖𝑖2 ⊗ … ⊗ 𝜎𝜎𝑖𝑖𝑛𝑛 � ∝ (𝜎𝜎0 ⊗ 𝜎𝜎0 ⊗ … ⊗ 𝜎𝜎0 )

𝑖𝑖1 =0 𝑖𝑖2 =0 𝑖𝑖𝑛𝑛 =0

And decide the normalization factor to make the proportionality become equality.

Answer) Similar to 5-3),

3 3 3

� � … � �𝜎𝜎𝑖𝑖1 ⊗ 𝜎𝜎𝑖𝑖2 ⊗ … ⊗ 𝜎𝜎𝑖𝑖𝑛𝑛 �𝑀𝑀�𝜎𝜎𝑖𝑖1 ⊗ 𝜎𝜎𝑖𝑖2 ⊗ … ⊗ 𝜎𝜎𝑖𝑖𝑛𝑛 �

𝑖𝑖1 =0 𝑖𝑖2 =0 𝑖𝑖𝑛𝑛 =0

= 2𝑛𝑛 𝑇𝑇𝑇𝑇[𝑀𝑀](𝜎𝜎0 ⊗ 𝜎𝜎0 ⊗ … ⊗ 𝜎𝜎0 ) ∝ (𝜎𝜎0 ⊗ 𝜎𝜎0 ⊗ … ⊗ 𝜎𝜎0 )

6) Find an orthogonal basis of ℂ𝑑𝑑×𝑑𝑑 with respect to the Hilbert-Schmidt norm.

Answer) If there exists an orthogonal basis of ℂ𝑑𝑑×𝑑𝑑 with respect to the Hilbert-Schmidt
norm, then the basis set {𝑀𝑀1 , 𝑀𝑀2 , … , 𝑀𝑀𝑑𝑑2 } should satisfy the condition 𝑇𝑇𝑇𝑇�𝑀𝑀𝑖𝑖† 𝑀𝑀𝑗𝑗 � ∝ 𝛿𝛿𝑖𝑖𝑖𝑖 .
6.

1) Write down the states obtained when Alice applies Pauli operators 𝐼𝐼, 𝑍𝑍, 𝑋𝑋, and 𝑋𝑋𝑋𝑋 on
the given Bell state, respectively.

Answer) Bob’s qubit does not affected when Alice applies Pauli operators on the given
Bell state.
1
⇒ 𝐼𝐼𝐴𝐴 ⊗ 𝐼𝐼𝐵𝐵 |Φ+ ⟩𝐴𝐴𝐴𝐴 = (|00⟩𝐴𝐴𝐴𝐴 + |11⟩𝐴𝐴𝐴𝐴 ) ≡ |𝜓𝜓1 ⟩
√2
1
𝑍𝑍𝐴𝐴 ⊗ 𝐼𝐼𝐵𝐵 |Φ+ ⟩𝐴𝐴𝐴𝐴 = (|00⟩𝐴𝐴𝐴𝐴 − |11⟩𝐴𝐴𝐴𝐴 ) ≡ |𝜓𝜓2 ⟩
√2
1
𝑋𝑋𝐴𝐴 ⊗ 𝐼𝐼𝐵𝐵 |Φ+ ⟩𝐴𝐴𝐴𝐴 = (|10⟩𝐴𝐴𝐴𝐴 + |01⟩𝐴𝐴𝐴𝐴 ) ≡ |𝜓𝜓3 ⟩
√2
1
𝑋𝑋𝐴𝐴 𝑍𝑍𝐴𝐴 ⊗ 𝐼𝐼𝐵𝐵 |Φ+ ⟩𝐴𝐴𝐴𝐴 = (|10⟩𝐴𝐴𝐴𝐴 − |01⟩𝐴𝐴𝐴𝐴 ) ≡ |𝜓𝜓4 ⟩
√2

2) Prove that the four states obtained in 1) are orthogonal to each other.

Answer) Skip since it is too simple calculation.

3) Construct a set of measurement operators that can perfectly distinguish the four
states with no failure.

Answer) If we construct a set with projection operators of each state, then

4

�|𝜓𝜓𝑖𝑖 ⟩⟨𝜓𝜓𝑖𝑖 | = |00⟩𝐴𝐴𝐴𝐴 ⟨00|𝐴𝐴𝐴𝐴 + |01⟩𝐴𝐴𝐴𝐴 ⟨01|𝐴𝐴𝐴𝐴 + |10⟩𝐴𝐴𝐴𝐴 ⟨10|𝐴𝐴𝐴𝐴 + |11⟩𝐴𝐴𝐴𝐴 ⟨11|𝐴𝐴𝐴𝐴 = 𝐼𝐼
𝑖𝑖=1

Also, since the states are orthogonal, they can perfectly distinguish.

4) Suppose there is a malevolent third party, Eve, and she intercepts the qubit Alice
sends. Prove that Eve cannot get any information about what Alice wanted to send using
the intercepted qubit.

Answer) Eve can only get the partial trace of the state. Since the Bell state is maximally
entangled, Eve cannot get any information from the intercepted qubit because it gives
totally random result.
7.

1) Prove that the measurement in Bell basis (Bell measurement) forms a proper
measurement set.

Answer) It is completely the same with 6-3).

2) Suppose Alice measures her qubits of 𝐴𝐴0 , 𝐴𝐴 systems using the projectors of the Bell
basis. Find the probabilities of obtaining each outcome and the post-measurement state
of Bob’s qubit for each outcome.

Answer) Suppose the qubit |𝜓𝜓⟩𝐴𝐴0 = 𝛼𝛼|0⟩ + 𝛽𝛽|1⟩ where |𝛼𝛼|2 + |𝛽𝛽|2 = 1.
1
Then the composed qubit state |𝜓𝜓⟩𝐴𝐴0 |Φ+ ⟩𝐴𝐴𝐴𝐴 = [𝛼𝛼|0⟩ + 𝛽𝛽|1⟩]𝐴𝐴0 [|00⟩𝐴𝐴𝐴𝐴 + |11⟩𝐴𝐴𝐴𝐴 ]
√2

1
= �|Φ+ ⟩𝐴𝐴0 𝐴𝐴 (𝛼𝛼|0⟩ + 𝛽𝛽|1⟩)𝐵𝐵 + |Φ− ⟩𝐴𝐴0 𝐴𝐴 (𝛼𝛼|0⟩ − 𝛽𝛽|1⟩)𝐵𝐵 + |Ψ+ ⟩𝐴𝐴0 𝐴𝐴 (𝛽𝛽|0⟩ + 𝛼𝛼|1⟩)𝐵𝐵
2
+ |Φ− ⟩𝐴𝐴0 𝐴𝐴 (𝛽𝛽|0⟩ − 𝛼𝛼|1⟩)𝐵𝐵 �

∴ Each outcome has the same probability, and the post-measurement state of Bob’s
qubit for each outcome is shown.

3) Construct unitary operators with which Bob can recover the initial state of Alice |𝜓𝜓⟩
for each case.

Answer) If Alice measures |Φ+ ⟩ state, Bob should apply 𝐼𝐼 operator(=already initial state).

If Alice measures |Φ− ⟩ state, Bob should apply 𝑍𝑍 operator.

If Alice measures |Ψ+ ⟩ state, Bob should apply 𝑋𝑋 operator.

If Alice measures |Ψ− ⟩ state, Bob should apply 𝑋𝑋𝑋𝑋 operator.

8.

1) Construct a unitary operation 𝑈𝑈 that transforms 𝑑𝑑 orthogonal states {|𝜓𝜓𝑖𝑖 ⟩}𝑑𝑑𝑖𝑖=1 such
that 𝑈𝑈|𝜓𝜓𝑖𝑖 ⟩|0⟩ = |𝜓𝜓𝑖𝑖 ⟩|𝜓𝜓𝑖𝑖 ⟩, where |0⟩ is a state in the Hilbert space.

Answer) We can construct d+1-dimensional CNOT gate i.e. swap the state |0⟩ and |𝜓𝜓𝑖𝑖 ⟩
when the initial qubit is in |𝜓𝜓𝑖𝑖 ⟩ state, and otherwise let the state at rest. This CNOT gate
is obviously unitary, and can clone a orthogonal state.

2) Prove that there is no unitary operation that can clone a quantum state. Nevertheless,
there is a unitary that can clone orthogonal states.

Answer) Let 𝑈𝑈|𝜓𝜓𝑖𝑖 ⟩|0⟩ = |𝜓𝜓𝑖𝑖 ⟩|𝜓𝜓𝑖𝑖 ⟩ and 𝑈𝑈�𝜓𝜓𝑗𝑗 �|0⟩ = �𝜓𝜓𝑗𝑗 ��𝜓𝜓𝑗𝑗 �.

Then �𝜓𝜓𝑖𝑖 �𝜓𝜓𝑗𝑗 � = (⟨𝜓𝜓𝑖𝑖 | ⊗ ⟨0|)��𝜓𝜓𝑗𝑗 � ⊗ |0⟩�

2
= (⟨𝜓𝜓𝑖𝑖 | ⊗ ⟨0|)𝑈𝑈 † 𝑈𝑈��𝜓𝜓𝑗𝑗 � ⊗ |0⟩� = ��𝜓𝜓𝑖𝑖 �𝜓𝜓𝑗𝑗 ��

Which means, �𝜓𝜓𝑖𝑖 �𝜓𝜓𝑗𝑗 � =0 or 1, namely |𝜓𝜓𝑖𝑖 ⟩ and �𝜓𝜓𝑗𝑗 � are identical or orthogonal.

∴ One can only clone states which are orthogonal to one another.

9.

1) Prove the HJW theorem: The sets of ��𝜓𝜓�𝑖𝑖 ��𝑖𝑖 and ��𝜙𝜙�𝑗𝑗 ��𝑗𝑗 , where �𝜓𝜓�𝑖𝑖 � and �𝜙𝜙�𝑗𝑗 � may not
be normalized, generate the same density matrix, i.e., ∑𝑖𝑖�𝜓𝜓�𝑖𝑖 ��𝜓𝜓�𝑖𝑖 � = ∑𝑗𝑗�𝜙𝜙�𝑗𝑗 ��𝜙𝜙�𝑗𝑗 �, if and only
if �𝜓𝜓�𝑖𝑖 � = ∑𝑗𝑗 𝑢𝑢𝑖𝑖𝑖𝑖 �𝜙𝜙�𝑗𝑗 �, where 𝑢𝑢 is a unitary matrix. When the size of the sets is not equal to
each other, you may introduce more vectors to a smaller set.

Answer)
† �
(1) If �𝜓𝜓�𝑖𝑖 � = ∑𝑗𝑗 𝑢𝑢𝑖𝑖𝑖𝑖 �𝜙𝜙�𝑗𝑗 �, then ∑𝑖𝑖�𝜓𝜓�𝑖𝑖 ��𝜓𝜓�𝑖𝑖 � = ∑𝑖𝑖𝑗𝑗𝑗𝑗 𝑢𝑢𝑖𝑖𝑖𝑖 𝑢𝑢𝑘𝑘𝑘𝑘 �𝜙𝜙𝑗𝑗 ��𝜙𝜙�𝑘𝑘 �
Here, ∑𝑗𝑗𝑗𝑗 𝑢𝑢𝑖𝑖𝑖𝑖 𝑢𝑢† = 𝛿𝛿𝑘𝑘𝑘𝑘 ⇒ ∑𝑖𝑖𝑗𝑗𝑗𝑗 𝑢𝑢𝑖𝑖𝑖𝑖 𝑢𝑢† �𝜙𝜙�𝑗𝑗 ��𝜙𝜙�𝑘𝑘 � = ∑𝑗𝑗�𝜙𝜙�𝑗𝑗 ��𝜙𝜙�𝑗𝑗 �
𝑘𝑘𝑘𝑘 𝑘𝑘𝑘𝑘
(2) Let the linear transformation as �𝜓𝜓�𝑖𝑖 � = ∑𝑗𝑗 𝑀𝑀𝑖𝑖𝑖𝑖 �𝜙𝜙�𝑗𝑗 �, then
∑𝑖𝑖�𝜓𝜓�𝑖𝑖 ��𝜓𝜓�𝑖𝑖 � = ∑𝑖𝑖𝑗𝑗𝑗𝑗 𝑀𝑀𝑖𝑖𝑖𝑖 �𝜙𝜙�𝑗𝑗 ��𝜓𝜓�𝑘𝑘 �𝑀𝑀𝑖𝑖𝑖𝑖
∗
= ∑𝑖𝑖𝑗𝑗𝑗𝑗�𝑀𝑀† 𝑀𝑀𝑖𝑖𝑖𝑖 ��𝜙𝜙�𝑗𝑗 ��𝜓𝜓�𝑘𝑘 � �= ∑𝑖𝑖�𝜙𝜙�𝑖𝑖 ��𝜙𝜙�𝑖𝑖 ��.
𝑘𝑘𝑘𝑘
†
Thus, ∑𝑖𝑖�𝑀𝑀𝑘𝑘𝑘𝑘 𝑀𝑀𝑖𝑖𝑖𝑖 � = 𝛿𝛿𝑘𝑘𝑘𝑘 , which means that 𝑀𝑀𝑖𝑖𝑖𝑖 is an unitary matrix.

2) Suppose we have two composite states |𝜓𝜓𝐴𝐴𝐴𝐴 ⟩ and |𝜙𝜙𝐴𝐴𝐴𝐴 ⟩ and they have the same
Schmidt coefficients. Prove that they are related by

|𝜓𝜓𝐴𝐴𝐴𝐴 ⟩ = (𝑈𝑈𝐴𝐴 ⊗ 𝑈𝑈𝐵𝐵 )|𝜙𝜙𝐴𝐴𝐴𝐴 ⟩

For some unitary operators 𝑈𝑈𝐴𝐴 and 𝑈𝑈𝐵𝐵 .

Answer) Let |𝜓𝜓𝐴𝐴𝐴𝐴 ⟩ = ∑𝑖𝑖 �𝜆𝜆𝑖𝑖 �𝜓𝜓𝐴𝐴 𝑖𝑖 � ⊗ �𝜓𝜓𝐵𝐵𝑖𝑖 � and |𝜙𝜙𝐴𝐴𝐴𝐴 ⟩ = ∑𝑗𝑗 �𝜆𝜆𝑗𝑗 �𝜙𝜙𝐴𝐴 𝑗𝑗 � ⊗ �𝜙𝜙𝐵𝐵𝑗𝑗 �.
Define the operator 𝑈𝑈𝐴𝐴 = ∑𝑖𝑖�𝜓𝜓𝐴𝐴 𝑖𝑖 � �𝜙𝜙𝐴𝐴𝑗𝑗 � and 𝑉𝑉𝐵𝐵 = ∑𝑖𝑖�𝜓𝜓𝐵𝐵 𝑖𝑖 � �𝜙𝜙𝐵𝐵𝑗𝑗 �, then

(𝑈𝑈𝐴𝐴 ⊗ 𝑈𝑈𝐵𝐵 )|𝜙𝜙𝐴𝐴𝐴𝐴 ⟩ = �� �𝜓𝜓𝐴𝐴 𝑗𝑗 � �𝜙𝜙𝐴𝐴𝑗𝑗 � ⊗ �𝜓𝜓𝐵𝐵 𝑘𝑘 ��𝜙𝜙𝐵𝐵𝑘𝑘 �� � �𝜆𝜆𝑖𝑖 �𝜙𝜙𝐴𝐴 𝑖𝑖 � ⊗ �𝜙𝜙𝐵𝐵𝑖𝑖 �
𝑗𝑗,𝑘𝑘 𝑖𝑖

= � �𝜆𝜆𝑖𝑖 �𝜙𝜙𝐴𝐴𝑗𝑗 �𝜙𝜙𝐴𝐴𝑖𝑖 � �𝜓𝜓𝐴𝐴𝑘𝑘 � ⊗ �𝜙𝜙𝐵𝐵𝑘𝑘 �𝜙𝜙𝐵𝐵𝑖𝑖 ��𝜓𝜓𝐵𝐵𝑘𝑘 � = � �𝜆𝜆𝑘𝑘 �𝜓𝜓𝐴𝐴𝑘𝑘 � ⊗ �𝜓𝜓𝐵𝐵𝑘𝑘 � = |𝜓𝜓𝐴𝐴𝐴𝐴 ⟩
𝑖𝑖𝑗𝑗𝑗𝑗 𝑘𝑘

3) Suppose we have a density matrix 𝜌𝜌𝐴𝐴 and we have its purifications |𝐴𝐴𝑅𝑅1 ⟩ and |𝐴𝐴𝑅𝑅2 ⟩,
i.e., their reduced density matrices for 𝐴𝐴 part are 𝜌𝜌𝐴𝐴 . Prove that the two states are
related by |𝐴𝐴𝑅𝑅1 ⟩ = (𝐼𝐼𝐴𝐴 ⊗ 𝑈𝑈𝐵𝐵 )|𝐴𝐴𝑅𝑅2 ⟩ for some unitary operator 𝑈𝑈𝐵𝐵 on 𝐵𝐵 part.

Answer) We can decompose 𝜌𝜌𝐴𝐴 as 𝜌𝜌𝐴𝐴 = ∑𝑖𝑖 𝑝𝑝𝑖𝑖 |𝜓𝜓𝐴𝐴 ⟩⟨𝜓𝜓𝐴𝐴 |, and let the system 𝐵𝐵 having
orthonormal basis as �𝜙𝜙𝑅𝑅1 ��𝜙𝜙𝑅𝑅1 �, then |𝐴𝐴𝑅𝑅1 ⟩ = ∑𝑖𝑖 𝑝𝑝𝑖𝑖 |𝜓𝜓𝐴𝐴 ⟩�𝜙𝜙𝑅𝑅1 �.

Obviously, orthonormal basis can be transformed into another orthonormal basis with
unitary transformation. Thus, we can choose a proper unitary transformation that satisfies
|𝐴𝐴𝑅𝑅1 ⟩ = (𝐼𝐼𝐴𝐴 ⊗ 𝑈𝑈𝐵𝐵 )|𝐴𝐴𝑅𝑅2 ⟩.

10.

1) Show an explicit construct of such a circuit.

Answer) One can find in the lecture notes.

2) Generalize the construction to 𝑓𝑓 ∶ {0,1}𝑚𝑚 → {0,1}𝑛𝑛 , where the output is now 𝑛𝑛 bits.

Answer) Again, we may use the induction.

 (1) There exist 𝑓𝑓 ∶ {0,1}𝑚𝑚 → {0,1} from 10-1).
(2) Suppose there exists the function 𝑓𝑓 ∶ {0,1}𝑚𝑚 → {0,1}𝑛𝑛 . Then we can construct
𝑓𝑓 ∶ {0,1}𝑚𝑚 → {0,1}𝑛𝑛+1 by using the function 𝑓𝑓 ∶ {0,1}𝑚𝑚 → {0,1} for additional
results.

From these two propositions, we can construct 𝑓𝑓 ∶ {0,1}𝑚𝑚 → {0,1}𝑛𝑛 .

3) Prove that Toffoli gate is universal, i.e., any Boolean function can be implemented by
using Toffoli gates.

Answer) Any Boolean function can be constructed from “AND”, “OR”, “NOT”. Or, there are
several universal gates (for example NAND gate). If you properly construct those three
gates or some universal gates, I may give you points.