Taking a closer look at Quantum gates and their operations

This post is a continuation of my earlier post ‘Exploring Quantum gate operations with QCSimulator’. Here I take a closer look at more quantum gates and their operations, besides implementing these new gates in my Quantum Computing simulator, the  QCSimulator in R.

Disclaimer: This article represents the author’s viewpoint only and doesn’t necessarily represent IBM’s positions, strategies or opinions

In  quantum circuits, gates  are unitary matrices which operate on 1,2 or 3 qubit systems which are represented as below

1 qubit
|0> = \begin{pmatrix}1\\0\end{pmatrix} and |1> = \begin{pmatrix}0\\1\end{pmatrix}

2 qubits
|0> \otimes |0> = \begin{pmatrix}1\\ 0\\ 0\\0\end{pmatrix}
|0> \otimes |1> = \begin{pmatrix}0\\ 1\\ 0\\0\end{pmatrix}
|1> \otimes |o> = \begin{pmatrix}0\\ 0\\ 1\\0\end{pmatrix}
|1> \otimes |1> = \begin{pmatrix}0\\ 0\\ 0\\1\end{pmatrix}

3 qubits
|0> \otimes |0> \otimes |0> = \begin{pmatrix}1\\ 0\\0\\ 0\\ 0\\0\\ 0\\0\end{pmatrix}
|0> \otimes |0> \otimes |1> = \begin{pmatrix}0\\ 1\\0\\ 0\\ 0\\0\\ 0\\0\end{pmatrix}
|0> \otimes |1> \otimes |0> = \begin{pmatrix}0\\ 0\\1\\ 0\\ 0\\0\\ 0\\0\end{pmatrix}


|1> \otimes |1> \otimes |1> = \begin{pmatrix}0\\ 0\\0\\ 0\\ 0\\0\\ 0\\1\end{pmatrix}
Hence single qubit is represented as 2 x 1 matrix, 2 qubit as 4 x 1 matrix and 3 qubit as 8 x 1 matrix

1) Composing Quantum gates in series
When quantum gates are connected in a series. The overall effect of the these quantum gates in series is obtained my taking the dot product of the unitary gates in reverse. For e.g.
Untitled

In the following picture the effect of the quantum gates A,B,C is the dot product of the gates taken reverse order
result = C . B . A

This overall action of the 3 quantum gates can be represented by a single ‘transfer’ matrix which is the dot product of the gates
Untitled

If we had a Pauli X followed by a Hadamard gate the combined effect of these gates on the inputs can be deduced by constructing a truth table

Input Pauli X – Output A’ Hadamard – Output B
|0> |1> 1/√2(|0>  – |1>)
|1> |0> 1/√2(|0>  + |1>)

Or

|0> -> 1/√2(|0>  – |1>)
|1> -> 1/√2(|0>  + |1>)
which is
\begin{pmatrix}1\\0\end{pmatrix}  ->1/√2 \begin{pmatrix}1\\0\end{pmatrix}\begin{pmatrix}0\\1\end{pmatrix} = 1/√2  \begin{pmatrix}1\\-1\end{pmatrix}
\begin{pmatrix}0\\1\end{pmatrix}  ->1/√2 \begin{pmatrix}1\\0\end{pmatrix} + \begin{pmatrix}0\\1\end{pmatrix} = 1/√2  \begin{pmatrix}1\\1\end{pmatrix}
Therefore the ‘transfer’ matrix can be written as
T = 1/√2 \begin{pmatrix}1 & 1\\ -1 & 1\end{pmatrix}

2)Quantum gates in parallel
When quantum gates are in parallel then the composite effect of the gates can be obtained by taking the tensor product of the quantum gates.
Untitled

If we consider the combined action of a Pauli X gate and a Hadamard gate in parallel
Untitled

A B A’ B’
|0> |0> |1> 1/√2(|0>  + |1>)
|0> |1> |1> 1/√2(|0>  – |1>)
|1> |0> |0> 1/√2(|0>  + |1>)
|1> |1> |0> 1/√2(|0>  – |1>)

Or

|00> => |1> \otimes 1/√2(|0>  + |1>) = 1/√2 (|10> + |11>)
|01> => |1> \otimes 1/√2(|0>  – |1>) = 1/√2 (|10> – |11>)
|10> => |0> \otimes 1/√2(|0>  + |1>) = 1/√2 (|00> + |01>)
|11> => |0> \otimes 1/√2(|0>  – |1>) = 1/√2 (|10> – |11>)

|00> = \begin{pmatrix}1\\ 0\\ 0\\0\end{pmatrix} =>1/√2\begin{pmatrix} 0\\ 0\\ 1\\ 1\end{pmatrix}
|01> = \begin{pmatrix}0\\ 1\\ 0\\0\end{pmatrix} =>1/√2\begin{pmatrix} 0\\ 0\\ 1\\ -1\end{pmatrix}
|10> = \begin{pmatrix}0\\ 0\\ 1\\0\end{pmatrix} =>1/√2\begin{pmatrix} 1\\ 0\\ 1\\ -1\end{pmatrix}
|11> = \begin{pmatrix}0\\ 0\\ 0\\1\end{pmatrix} =>1/√2\begin{pmatrix} 1\\ 0\\ -1\\ -1\end{pmatrix}

Here are more Quantum gates
a) Rotation gate
U = \begin{pmatrix}cos\theta & -sin\theta\\ sin\theta & cos\theta\end{pmatrix}

b) Toffoli gate
The Toffoli gate flips the 3rd qubit if the 1st and 2nd qubit are |1>

Toffoli gate
Input Output
|000> |000>
|001> |001>
|010> |010>
|011> |011>
|100> |100>
|101> |101>
|110> |111>
|111> |110>

c) Fredkin gate
The Fredkin gate swaps the 2nd and 3rd qubits if the 1st qubit is |1>

Fredkin gate
Input Output
|000> |000>
|001> |001>
|010> |010>
|011> |011>
|100> |100>
|101> |110>
|110> |101>
|111> |111>

d) Controlled U gate
A controlled U gate can be represented as
controlled U = \begin{pmatrix}1 & 0 & 0 & 0\\ 0 &1  &0  & 0\\ 0 &0  &u11  &u12 \\ 0 & 0 &u21  &u22 \end{pmatrix}   – (A)
where U =  \begin{pmatrix}u11 &u12 \\ u21 & u22\end{pmatrix}

e) Controlled Pauli gates
Controlled Pauli gates are created based on the following identities. The CNOT gate is a controlled Pauli X gate where controlled U is a Pauli X gate and matches the CNOT unitary matrix. Pauli gates can be constructed using

a) H x X x H = Z    &
H x H = I

b) S x X x S1
S x S1 = I

the controlled Pauli X, Y , Z are contructed using the CNOT for the controlled X in the above identities
In general a controlled Pauli gate can be created as below
Untitled

f) CPauliX
Here C is the 2 x2  Identity matrix. Simulating this in my QCSimulator
CPauliX I=matrix(c(1,0,0,1),nrow=2,ncol=2)
# Compute 1st composite
a = TensorProd(I,I)
b = CNOT2_01(a)
# Compute 1st composite
c = TensorProd(I,I)
#Take dot product
d = DotProduct(c,b)
#Take dot product with qubit
e = DotProduct(d,q)
e
}

Implementing the above with I, S, H gives Pauli X, Y and Z as seen below

library(QCSimulator)
I4=matrix(c(1,0,0,0,
            0,1,0,0,
            0,0,1,0,
            0,0,0,1),nrow=4,ncol=4)

#Controlled Pauli X
CPauliX(I4)
##      [,1] [,2] [,3] [,4]
## [1,]    1    0    0    0
## [2,]    0    1    0    0
## [3,]    0    0    0    1
## [4,]    0    0    1    0
#Controlled Pauli Y
CPauliY(I4)
##      [,1] [,2] [,3] [,4]
## [1,] 1+0i 0+0i 0+0i 0+0i
## [2,] 0+0i 1+0i 0+0i 0+0i
## [3,] 0+0i 0+0i 0+0i 0-1i
## [4,] 0+0i 0+0i 0+1i 0+0i
#Controlled Pauli Z
CPauliZ(I4)
##      [,1] [,2] [,3] [,4]
## [1,]    1    0    0    0
## [2,]    0    1    0    0
## [3,]    0    0    1    0
## [4,]    0    0    0   -1

g) CSWAP gate

Untitled

q00=matrix(c(1,0,0,0),nrow=4,ncol=1)
q01=matrix(c(0,1,0,0),nrow=4,ncol=1)
q10=matrix(c(0,0,1,0),nrow=4,ncol=1)
q11=matrix(c(0,0,0,1),nrow=4,ncol=1)
CSWAP(q00)
##      [,1]
## [1,]    1
## [2,]    0
## [3,]    0
## [4,]    0
#Swap qubits 
CSWAP(q01)
##      [,1]
## [1,]    0
## [2,]    0
## [3,]    1
## [4,]    0
#Swap qubits 
CSWAP(q10)
##      [,1]
## [1,]    0
## [2,]    1
## [3,]    0
## [4,]    0
CSWAP(q11)
##      [,1]
## [1,]    0
## [2,]    0
## [3,]    0
## [4,]    1

h) Toffoli state
The Toffoli state creates a 3 qubit entangled state 1/2(|000> + |001> + |100> + |111>)
Untitled

Simulating the Toffoli state in IBM Quantum Experience we get
Untitled

h) Implementation of Toffoli state in QCSimulator 

#ToffoliState 
    # Computation of the Toffoli State
    H=1/sqrt(2) * matrix(c(1,1,1,-1),nrow=2,ncol=2)
    I=matrix(c(1,0,0,1),nrow=2,ncol=2)

    # 1st composite
    # H x H x H
    a = TensorProd(TensorProd(H,H),H)
    # 1st CNOT
    a1= CNOT3_12(a)

    # 2nd composite
    # I x I x T1Gate
    b = TensorProd(TensorProd(I,I),T1Gate(I))
    b1 = DotProduct(b,a1)
    c = CNOT3_02(b1)

    # 3rd composite
    # I x I x TGate
    d = TensorProd(TensorProd(I,I),TGate(I))
    d1 = DotProduct(d,c)
    e = CNOT3_12(d1)

    # 4th composite
    # I x I x T1Gate
    f = TensorProd(TensorProd(I,I),T1Gate(I))
    f1 = DotProduct(f,e)
    g = CNOT3_02(f1)

    #5th composite
    # I x T x T
    h = TensorProd(TensorProd(I,TGate(I)),TGate(I))
    h1 = DotProduct(h,g)
    i = CNOT3_12(h1)

    #6th composite
    # I x H x H
    j = TensorProd(TensorProd(I,Hadamard(I)),Hadamard(I))
    j1 = DotProduct(j,i)
    k = CNOT3_12(j1)

    # 7th composite
    # I x H x H
    l = TensorProd(TensorProd(I,Hadamard(I)),Hadamard(I))
    l1 = DotProduct(l,k)
    m = CNOT3_12(l1)
    n = CNOT3_02(m)

    #8th composite
    # T x H x T1
    o = TensorProd(TensorProd(TGate(I),Hadamard(I)),T1Gate(I))
    o1 = DotProduct(o,n)
    p = CNOT3_02(o1)
    result = measurement(p)
    plotMeasurement(result)

a-1
The measurement is identical to the that of IBM Quantum Experience

Conclusion:  This post looked at more Quantum gates. I have implemented all the gates in my QCSimulator which I hope to release in a couple of months.

Disclaimer: This article represents the author’s viewpoint only and doesn’t necessarily represent IBM’s positions, strategies or opinions

References
1. http://www1.gantep.edu.tr/~koc/qc/chapter4.pdf
2. http://iontrap.umd.edu/wp-content/uploads/2016/01/Quantum-Gates-c2.pdf
3. https://quantumexperience.ng.bluemix.net/

Also see
1.  Venturing into IBM’s Quantum Experience
2. Going deeper into IBM’s Quantum Experience!
3.  A primer on Qubits, Quantum gates and Quantum Operations
4. Exploring Quantum gate operations with QCSimulator

Take a look at my other posts at
1. Index of posts

Exploring Quantum Gate operations with QCSimulator

Introduction: Ever since I was initiated into Quantum Computing, through IBM’s Quantum Experience I have been hooked. My initial encounter with domain made me very excited and all wound up. The reason behind this, I think, is because there is an air of mystery around ‘Quantum’ anything.  After my early rush with the Quantum Experience, I have deliberately slowed down to digest the heady stuff.

This post also includes my early prototype of a Quantum Computing Simulator( QCSimulator) which I am creating in R. I hope to have a decent Quantum Computing simulator available, in the next couple of months. The ideas for this simulator are based on IBM’s Quantum Experience and the lectures on Quantum Mechanics and Quantum Computation by Prof Umesh Vazirani from University of California at Berkeley at edX. This calls to this simulator have been included in R Markdown file and has been published at RPubs as Quantum Computing Simulator

Disclaimer: This article represents the author’s viewpoint only and doesn’t necessarily represent IBM’s positions, strategies or opinions

In this post I explore quantum gate operations

A) Quantum Gates
Quantum gates are represented as a n x n unitary matrix. In mathematics, a complex square matrix U is unitary if its conjugate transpose Uǂ is also its inverse – that is, if
U ǂU =U U ǂ=I

a) Clifford Gates
The following gates are known as Clifford Gates and are represented as the unitary matrix below

1. Pauli X
\begin{pmatrix}0&1\\1&0\end{pmatrix}

2.Pauli Y
\begin{pmatrix}0&-i\\i&0\end{pmatrix}

3. Pauli Z
\begin{pmatrix}1&0\\0&-1\end{pmatrix}

4. Hadamard
1/√2 \begin{pmatrix}1 & 1\\ 1 & -1\end{pmatrix}

5. S Gate
\begin{pmatrix}1 & 0\\ 0 & i\end{pmatrix}

6. S1 Gate
\begin{pmatrix}1 & 0\\ 0 & -i\end{pmatrix}

7. CNOT
\begin{pmatrix}1 & 0 & 0 &0 \\ 0 & 1 & 0 &0 \\  0& 0 &  0&1 \\ 0 & 0 & 1 & 0\end{pmatrix}

b) Non-Clifford Gate
The following are the non-Clifford gates
1. Toffoli Gate
T = \begin{pmatrix}1 & 0\\ 0 & e^{^{i\prod /4}}\end{pmatrix}

2. Toffoli1 Gate
T1 = \begin{pmatrix}1 & 0\\ 0 & e^{^{-i\prod /4}}\end{pmatrix}

B) Evolution of a 1 qubit Quantum System
The diagram below shows how a 1 qubit system evolves on the application of Quantum Gates.

Untitled

C) Evolution of a 2 Qubit  System
The following diagram depicts the evolution of a 2 qubit system. The 4 different maximally entangled states can be obtained by using a Hadamard and a CNOT gate to |00>, |01>, |10> & |11> resulting in the entangled states  Φ+, Ψ+, Φ, Ψrespectively

Untitled

D) Verifying Unitary’ness
XXǂ = XǂX= I
TTǂ = TǂT=I
SSǂ = SǂS=I
The Uǂ  function in the simulator is
Uǂ = GateDagger(U)

E) A look at some Simulator functions
The unitary functions for the Clifford and non-Clifford gates have been implemented functions. The unitary functions can be chained together by invoking each successive Gate as argument to the function.

1. Creating the dagger function
HDagger = GateDagger(Hadamard)
HDagger x Hadamard
TDagger = GateDagger(TGate)
TDagger x TGate

H
##           [,1]       [,2]
## [1,] 0.7071068  0.7071068
## [2,] 0.7071068 -0.7071068
HDagger = GateDagger(H)
HDagger
##           [,1]       [,2]
## [1,] 0.7071068  0.7071068
## [2,] 0.7071068 -0.7071068
HDagger %*% H
##      [,1] [,2]
## [1,]    1    0
## [2,]    0    1
T
##      [,1]                 [,2]
## [1,] 1+0i 0.0000000+0.0000000i
## [2,] 0+0i 0.7071068+0.7071068i
TDagger = GateDagger(T)
TDagger
##      [,1]                 [,2]
## [1,] 1+0i 0.0000000+0.0000000i
## [2,] 0+0i 0.7071068-0.7071068i
TDagger %*% T
##      [,1] [,2]
## [1,] 1+0i 0+0i
## [2,] 0+0i 1+0i

2. Angle between 2 vectors – Inner product
The angle between 2 vectors can be obtained by taking the inner product between the vectors

#1. a is the diagonal vector 1/2 |0> + 1/2 |1> and b = q0 = |0>
diagonal <-  matrix(c(1/sqrt(2),1/sqrt(2)),nrow=2,ncol=1)
q0=matrix(c(1,0),nrow=2,ncol=1)
innerProduct(diagonal,q0)
##      [,1]
## [1,]   45
#2. a = 1/2|0> + sqrt(3)/2|1> and  b= 1/sqrt(2) |0> + 1/sqrt(2) |1>
a = matrix(c(1/2,sqrt(3)/2),nrow=2,ncol=1)
b = matrix(c(1/sqrt(2),1/sqrt(2)),nrow=2,ncol=1)
innerProduct(a,b)
##      [,1]
## [1,]   15

3. Chaining Quantum Gates
For e.g.
H x q0
S x H x q0 == > SGate(Hadamard(q0))

Or
H x S x S x H x q0 == > Hadamard(SGate(SGate(Hadamard))))

# H x q0
Hadamard(q0)
##           [,1]
## [1,] 0.7071068
## [2,] 0.7071068
# S x H x q0
SGate(Hadamard(q0))
##                      [,1]
## [1,] 0.7071068+0.0000000i
## [2,] 0.0000000+0.7071068i
# H x S x S x H x q0
Hadamard(SGate(SGate(Hadamard(q0))))
##      [,1]
## [1,] 0+0i
## [2,] 1+0i
# S x T x H x T x H x q0
SGate(TGate(Hadamard(TGate(Hadamard(q0)))))
##                      [,1]
## [1,] 0.8535534+0.3535534i
## [2,] 0.1464466+0.3535534i

4. Measurement
The output of Quantum Gate operations can be measured with
measurement(a)
measurement(q0)
measurement(Hadamard(q0))
a=SGate(TGate(Hadamard(TGate(Hadamard(I)))))
measurement(a)
measurement(SGate(TGate(Hadamard(TGate(Hadamard(I))))))

measurement(q0)
##   0 1
## v 1 0
measurement(Hadamard(q0))
##     0   1
## v 0.5 0.5
a <- SGate(TGate(Hadamard(TGate(Hadamard(q0)))))
measurement(a)
##           0         1
## v 0.8535534 0.1464466

5. Plot the measurements

plotMeasurement(q1)

unnamed-chunk-5-1

plotMeasurement(Hadamard(q0))

unnamed-chunk-5-2

a = measurement(SGate(TGate(Hadamard(TGate(Hadamard(q0))))))
plotMeasurement(a)

unnamed-chunk-5-3

6. Using the QCSimulator for one of the Bell tests
Here I compute the following measurement of  Bell state ZW  with the QCSimulator
Untitled

When this is simulated on IBM’s Quantum Experience the result is
Untitled

Below I simulate the same on my R based QCSimulator

# Compute the effect of the Composite H gate with the Identity matrix (I)
a=kronecker(H,I,"*")
a
##           [,1]      [,2]       [,3]       [,4]
## [1,] 0.7071068 0.0000000  0.7071068  0.0000000
## [2,] 0.0000000 0.7071068  0.0000000  0.7071068
## [3,] 0.7071068 0.0000000 -0.7071068  0.0000000
## [4,] 0.0000000 0.7071068  0.0000000 -0.7071068
# Compute the applcation of CNOT on this result
b = CNOT(a)
b
##           [,1]      [,2]       [,3]       [,4]
## [1,] 0.7071068 0.0000000  0.7071068  0.0000000
## [2,] 0.0000000 0.7071068  0.0000000  0.7071068
## [3,] 0.0000000 0.7071068  0.0000000 -0.7071068
## [4,] 0.7071068 0.0000000 -0.7071068  0.0000000
# Obtain the result of CNOT on q00
c = b %*% q00
c
##           [,1]
## [1,] 0.7071068
## [2,] 0.0000000
## [3,] 0.0000000
## [4,] 0.7071068
# Compute the effect of the composite HxTxHxS gates and the Identity matrix(I) for measurement
d=Hadamard(TGate(Hadamard(SGate(I))))
e=kronecker(I, d,"*")

# Applying  the composite gate on the output 'c'
f = e %*% c
# Measure the output
g <- measurement(f)
g
##          00        01        10        11
## v 0.4267767 0.0732233 0.0732233 0.4267767
#Plot the measurement
plotMeasurement(g)

aaunnamed-chunk-6-1
which is exactly the result obtained with IBM’s Quantum Experience!

Conclusion : In  this post I dwell on 1 and 2-qubit quantum gates and explore their operation. I have started to construct a  R based Quantum Computing Simulator. I hope to have a reasonable simulator in the next couple of months. Let’s see.

Watch this space!

Disclaimer: This article represents the author’s viewpoint only and doesn’t necessarily represent IBM’s positions, strategies or opinions

References
1.  IBM Quantum Experience – User Guide
2. Quantum Mechanics and Quantum Conputing, UC Berkeley

Also see
1.  Venturing into IBM’s Quantum Experience
2. Going deeper into IBM’s Quantum Experience!
3.  A primer on Qubits, Quantum gates and Quantum Operations

You may also like
1.  My TEDx talk on the “Internet of Things”
2. Experiments with deblurring using OpenCV
3. TWS-5: Google’s Page Rank: Predicting the movements of a random web walker

For more posts see
Index of posts

Going deeper into IBM’s Quantum Experience!

Introduction

In this post I delve deeper into IBM’s Quantum Experience. As mentioned in my earlier post “Venturing into IBM’s Quantum Experience”, IBM, has opened up its Quantum computing environment, to the general public, as the Quantum Experience. The access to Quantum Experience is through IBM’s Platform as a Service (PaaS) offering, Bluemix™. Clearly this is a great engineering feat, which integrates the highly precise environment of Quantum Computing, where the qubits are maintained at 5 milliKelvin, and the IBM Bluemix PaaS environment on Softlayer. The Quantum Experience, is in fact Quantum Computing as a Service (QCaaS).  In my opinion, the Quantum Experience, provides a glimpse of tomorrow, today,

Disclaimer: This article represents the author’s viewpoint only and doesn’t necessarily represent IBM’s positions, strategies or opinions

Note: Also by the way, feel free to holler if you find anything incorrect or off the mark in my post. I am just getting started on quantum computing so there may be slip ups.

A) Bloch sphere

6In my earlier post the operations of the X, Y, Z, H, S, and S1 were measured using the standard or diagonal basis and the results were in probabilities of the qubit(s). However, the probabilities alone, in the standard basis, are not enough to specify a quantum state because it does not capture the phase of the superposition. A convenient representation for a qubit is the Bloch sphere.

The general state of a quantum two-level system can be written in the form

|ψ⟩=α|0⟩+β|1⟩,

Where α and β are complex numbers with |α|2+|β|2=1. This leads to the “canonical” parameterized form

|ψ⟩= cos Θ/2 |0> + e sin Θ/2 |1>                        (A)

in terms of only two real numbers θ and φ, with natural ranges 0 ≤ θ ≤ π and 0 ≤ φ ≤ 2π. These are the same as the polar angles in 3-dimensional spherical coordinates, and this leads to the representation of the state (1) as a point on a unit sphere called the Bloch sphere.

In the notation of (A), the state |0> is represented by the North pole, and the state |1> by the South pole. Note:  The states |0> and |1> , n the Bloch sphere are not orthogonal to each other. The states on the equator of the Bloch sphere, correspond to superpositions of |0> and |1> with equal weights (θ = π∕2), and different phases, parameterized by the azimuthal angle φ (the “longitude”)

In the picture below  Bloch  measurements  are performed on the operations on the qubits

1

The results of the Bloch measurements for the combination of  quantum gates are shown below

i) Quantum gate operations and Bloch measurements

4

ii) Quantum gate operations as Superposition operations

5

B) Classical vs Quantum computing
A classical computer that has N-bits has 2^{N}possible configurations. However, at any
one point in time, it can be in one, and only one of  2N  configurations. Interestingly, the quantum computer also takes in a n -bit number and outputs a n -bit number; but because of the superposition principle and the possibility of entanglement, the intermediate state is very different.

A system which had N different mutually exclusive states can be represented as |1>, |2>. . . |N> using the Dirac’s bra-ket notation

A pure quantum state is a superposition of all these states
Φ = α1 1> + α2 2> + …. + αN N>
To describe it requires complex numbers, giving a lot more room for maneuvering.

C) The CNOT gate
The CNOT gate or the Controlled-Not gate is an example of a two-qubit quantum gate. The CNOT gate’s action is to flip (apply a NOT or X gate to) the target qubit if the control qubit is 1; otherwise it does nothing. The CNOT plays the role of the classical XOR gate, but unlike the XOR, The CNOT gate is a two-output gate and is reversible It is represented by the matrix by the following 4 x 4 matrix

\begin{pmatrix}1 & 0 & 0 &0 \\ 0 & 1 & 0 &0 \\  0& 0 &  0&1 \\ 0 & 0 & 1 & 0\end{pmatrix}

The CNOT gate flips the target bit if the control bit is 1, otherwise it does nothing if it’s 0:

More specifically
CNOT|0>|b> = |0>|b>
CNOT|1>|b>= |1>|1 – b>

The operation of the CNOT gate can be elaborated as below
The 2-qubit basis states can be represented as four-dimensional vectors
|00> = \begin{pmatrix} 1& 0 & 0 & 0 \end{pmatrix}^{T}
|01> = \begin{pmatrix} 0& 1 & 0 & 0 \end{pmatrix}^{T}
|10> = \begin{pmatrix} 0& 0 & 1 & 0 \end{pmatrix}^{T}
|11> = \begin{pmatrix} 0& 0 & 0 & 1 \end{pmatrix}^{T}

For example, a quantum state may be expanded as a linear combination of this basis:
|ψ⟩=a|00⟩+b|01⟩+c|10⟩+d|11⟩

The CNOT matrix can  be applied as below
CNOT*|ψ⟩=CNOT*(a|00⟩+b|01⟩+c|10⟩+d|11⟩)
= a*CNOT*|00⟩+…+d*CNOT*|11⟩

where you perform standard matrix multiplication on the basis vectors to get:
CNOT*|ψ⟩=a|00⟩+b|01⟩+c|11⟩+d|10⟩

In other words, the CNOT gate has transformed
|10⟩↦|11⟩ and |11⟩↦|10⟩

i) CNOT operations in R code

# CNOT gate
cnot= matrix(c(1,0,0,0,0,1,0,0,0,0,0,1,0,0,1,0),nrow=4,ncol=4)
cnot
##      [,1] [,2] [,3] [,4]
## [1,]    1    0    0    0
## [2,]    0    1    0    0
## [3,]    0    0    0    1
## [4,]    0    0    1    0
#a. Qubit |00> 
q00=matrix(c(1,0,0,0),nrow=4,ncol=1)
q00
##      [,1]
## [1,]    1
## [2,]    0
## [3,]    0
## [4,]    0
# CNOT *q00 ==> q00
a <- cnot %*% q00
a
##      [,1]
## [1,]    1
## [2,]    0
## [3,]    0
## [4,]    0
#b.Qubit |01>
q01=matrix(c(0,1,0,0),nrow=4,ncol=1)
q01
##      [,1]
## [1,]    0
## [2,]    1
## [3,]    0
## [4,]    0
# CNOT *q01 ==> q01
a <- cnot %*% q01
a
##      [,1]
## [1,]    0
## [2,]    1
## [3,]    0
## [4,]    0
#c. Qubit |10>
q10=matrix(c(0,0,1,0),nrow=4,ncol=1)
q10
##      [,1]
## [1,]    0
## [2,]    0
## [3,]    1
## [4,]    0
# CNOT *q10 ==> q11
a <- cnot %*% q10
a
##      [,1]
## [1,]    0
## [2,]    0
## [3,]    0
## [4,]    1
#d. Qubit |11>
q11=matrix(c(0,0,0,1),nrow=4,ncol=1)
q11
##      [,1]
## [1,]    0
## [2,]    0
## [3,]    0
## [4,]    1
# CNOT *q11 ==> q10
a <- cnot %*% q11
a
##      [,1]
## [1,]    0
## [2,]    0
## [3,]    1
## [4,]    0

D) Non Clifford gates
The quantum gates discussed in my earlier post (Pauli X, Y, Z, H, S and S1) and the CNOT are members of a special group of gates known as the ‘Clifford group’. These gates can be simulated efficiently on a classical computer. Therefore, the Clifford group is not universal.  A finite set of gates that can approximate any arbitrary unitary matrix is known as a universal gate set. This is similar,  to how certain sets of classical logic gates, such as {AND, NOT}, are functionally complete and can be used to build any Boolean function ( I remember this axiom/fact from my Digital Electronics -101 class about 3  decades back!).

Adding almost any non-Clifford gate to single-qubit Clifford gates and CNOT gates makes the group universal which I presume can simulate any arbitrary unitary matrix.  The non-Clifford gates, discussed are the T and  Tǂ gates

These are given by

T = \begin{pmatrix}1 & 0\\ 0 & e^{^{i\prod /4}}\end{pmatrix}

Tǂ =\begin{pmatrix}1 & 0\\ 0 & e^{^{-i\prod /4}}\end{pmatrix}

i) T gate operations
The T gate makes it possible to reach  different points of the Bloch sphere.  By increasing the number of T-gates in the quantum circuit ( we start to cover the Bloch sphere more densely with states that can be reached

ii)  T Gates of depth 2 – Computational Measurement
9

Simulating in Composer
10

iii) Simulating in R Code
Measurement only gives the real part and does not provide info on phase

# T Gate 
T=matrix(c(1,0,0,exp(1i*pi/4)),nrow=2,ncol=2)
# Simulating T Gate depth-2 - Computational  measurement
a=S%*%T%*%H%*%T%*%H%*%q0
a
##                      [,1]
## [1,] 0.8535534+0.3535534i
## [2,] 0.1464466+0.3535534i

iv) 2 T Gates – Bloch  Measurement
7

Bloch measurement

8

v) Simulating T gate in R code
This gives the phase values as shown in the Bloch sphere

# Simulating T Gate depth-2 - Bloch measurement (use a diagonal basis H gate in front)
a=H%*%S%*%T%*%H%*%T%*%H%*%q0
a
##                [,1]
## [1,] 0.7071068+0.5i
## [2,] 0.5000000+0.0i

 

E) Quantum Entanglement – The case of ‘The Spooky action at a distance’
One of the infamous counter-intuitive ideas of quantum mechanics is that two systems that appear too far apart to influence each other can nevertheless behave in ways that, though individually random, are too strongly correlated to be described by any classical local theory.  For e.g. when the 1st qubit of a pair of “quantum  entangled” qubits are measured, this automatically determines the 2nd qubit, though the individual qubits may be separated by extremely large distances. It appears that the measurement of the first qubit cannot affect the 2nd qubit which is too far apart to be be influenced and also given the fact  that nothing can travel faster than the speed of light.  More specifically

“Suppose Alice has the first qubit of the pair, and Bob has the second. If Alice measures her qubit in the computational basis and gets outcome b ∈ {0, 1},  then the state collapses to |bb> . In other words the measurements and outcome of the 1st qubit determines the outcome of the  2nd qubit . How weird is that?

Similarly, if Alice measures her qubit in some other basis, this will collapse the joint state (including Bob’s qubit) to some state that depends on her measurement basis as well as its outcome. Somehow Alice’s action seems to have an instantaneous effect on Bob’s side—even if the two qubits are light-years apart!”

How weird is that!

Einstein, whose theory of relativity posits that information and causation cannot travel faster than the speed of light, was greatly troubled by this, . Einstein called such effects of entanglement “spooky action at a distance”.

In the 1960s, John Bell devised entanglement-based experiments whose behavior cannot be reproduced by any “local realist” theory where the implication of local and realist is given below

Locality: No information can travel faster than the speed of light. There is a hidden variable that defines all the correlations.

Realism:  All observables have a definite value independent of the measurement

i) Bell state measurements
The mathematical proof for the Bell tests  are quite abstract  and mostly escaped my grasp. I  hope to get my arms around this beast, in the weeks and months to come. However, I understood how to run the tests and perform the calculations which are included below.  I have executed the Bell Tests on

a) Ideal Quantum Processor (Simulator with ideal conditions)
b) Realistic Quantum Processor (Simulator with realistic conditions)
c) Real Quantum Processor. For this I used 8192 ‘shots’ repeats of the experiment

I finally calculate |C| for all 3 tests

The steps involved in calculating |C|
1.  Execute ZW, ZV, XW, XV
2. Calculate = P(00) + P(11) – P(01) – P(10)
3.  Finally |C| = ZW + ZV + XW – XV

Preparation of Qubit |00>
The qubits are in state |00> The H gate  takes the first qubit to the equal superposition

1/√2(|00> + |10>)  and the CNOT gate flips the second qubit if the first is excited, making the state  1/√2(|00> + |11>). This is the entangled state (commonly called a Bell state)

11

Simulating in Composer
This prepares the entangled state 1/√2(|00> + |11>)

12

It can be seen that the the qubits |00> and  |11> are created

1) Simulations on Ideal Quantum processor

a) Bell state ZW (Ideal) 
13
Simulation
24
P(00) = 0.427
P(01) = 0.073
P(10) =0.073
P(11) = 0.427

b) Bell state ZV (Ideal) 
14

Simulation

25
P(00) = 0.427
P(01) = 0.073
P(10) =0.073
P(11) = 0.427

c) Bell state XW (Ideal)
26

Measurement
27
P(00) = 0.427
P(01) = 0.073
P(10) =0.073
P(11) = 0.427

d) Bell state XV (Ideal)
28

Simulating Bell State XV in Composer
16
P(00) = 0.073
P(01) = 0.473
P(10) = 0.473
P(11) =0.73

Bell test measurement in Ideal Quantum Processor are given below
Untitled

For the Ideal Quantum Processor
|C) = 2.832

2) Simulations on the Realistic  Quantum Processor
The Bell tests above were simulated on Realistic Quantum Processor. The results are included below

Untitled

For the Realistic Quantum Processor
|C) = 2.523

3) Real IBM Quantum  Processor (8192 shots)
Finally the Bell Tests were executed in IBM’s Real Quantum Processor for 8192 shots, each requiring 5 standard units. The tests were queued, executed and the results sent by mail. The results are included below
a) Bell State ZW measurement (Real)
29

b) Bell state ZV measurement  (Real)
30
c) Bell State XW measurement (Real)
31
d) Bell state XV measurement (Real)

32

;

The results were tabulated and |C| computed. Bell test measurement in Real Quantum Processor are given below

Untitled

The Bell measurements on the Reak Quantum Processor is
|C) = 2.509

Conclusion
This post included details on the CNOT and the non-Clifford gates. The Bell tests were performed on all 3 processors Ideal, Realistic and Real Quantum Processors and in each case the |C| > 2. While I have been to execute the tests I will definitely have to spend more time understanding the nuances.

I hope to continue this journey into quantum computing the months to come. Watch this space!

Disclaimer: This article represents the author’s viewpoint only and doesn’t necessarily represent IBM’s positions, strategies or opinions

References
1.IBM Quantum Experience – User Guide
2.Quantum Computing lecture notes – Ronald De Wolf
3.Bloch sphere
4.Basic concepts in quantum computation
5.Physics Stack Exchange

You may also like
1.Revisiting crimes against women in India
2.Natural language processing: What would Shakespeare say?
3. My book ‘Practical Machine Learning with R and Python’ on Amazon
4. Introducing QCSimulator: A 5-qubit quantum computing simulator in R
5.Beaten by sheer pace – Cricket analytics with yorkr
6.A method for optimal bandwidth usage by auctioning available bandwidth using the OpenFlow protocol
7.TWS-4: Gossip protocol: Epidemics and rumors to the rescue
8.Bend it like Bluemix, MongoDB using Auto-scale – Part 1!
9.Fun simulation of a Chain in Android
10.Cricket analytics with cricketr in paperback and Kindle versions