Implementing Quantum Machine Learning With Qiskit

Introduction

Quantum Machine Learning (QML) merges quantum computing and machine learning by utilizing the key aspects of both disciplines. Quantum computing has the potential to solve complex problems efficiently, while machine learning's adaptive algorithms can learn from and make predictions on data. This post will dive into how to implement a basic quantum machine learning algorithm using IBM's open-source quantum computing framework, Qiskit.

Quantum States and Qubits

Before we begin, it's important to understand a bit about quantum states and qubits. Quantum computers use quantum bits or qubits, which can be both 0 and 1 simultaneously, thanks to superposition. The state of a qubit can be represented as a complex wave function, a linear combination of its basis states.

# import necessary libraries
from qiskit import QuantumCircuit, transpile
from qiskit.providers.aer import QasmSimulator
from qiskit.visualization import plot_histogram

# Create a Quantum Circuit acting on a quantum register of one qubit
circ = QuantumCircuit(1)

# Add a H gate on qubit 0, putting this qubit in superposition.
circ.h(0)

# Visualize the circuit:
print(circ)

Quantum Machine Learning with Qiskit

Let us implement a very basic binary classifier using Quantum Support Vector Machine (QSVM) algorithm provided by Qiskit's "aqua" components. For this blog, we will use the "ad hoc" dataset, which is arbitrarily defined and frequently used for examples and testing.

from qiskit.ml.datasets import ad_hoc_data
from qiskit.aqua.utils import split_dataset_to_data_and_labels, map_label_to_class_name
from qiskit.aqua import QuantumInstance
from qiskit.aqua.algorithms import QSVM
from qiskit.aqua.components.feature_maps import SecondOrderExpansion

# Setting up the dataset
feature_dim = 2
sample_Total, training_input, test_input, class_labels = ad_hoc_data(
    training_size=20, 
    test_size=10, 
    n=feature_dim, 
    gap=0.3, 
    plot_data=True
)

## Set up QSVM
feature_map = SecondOrderExpansion(feature_dimension=feature_dim, depth=2)
qsvm = QSVM(feature_map, training_input, test_input)

# Set up the backend for the quantum instance
backend = BasicAer.get_backend('qasm_simulator')
quantum_instance = QuantumInstance(backend, shots=1024)

result = qsvm.run(quantum_instance)

print("testing success ratio: ", result['testing_accuracy'])

In the code above, we first set up the dataset, defining the dimension of the features. Then, we call the ad_hoc_data method from Qiskit's ml_datasets to generate our training and test datasets. After setting up our QSVM with a SecondOrderExpansion feature map, we define our quantum instance and run the QSVM with this instance. The QSVM.run() method returns a dictionary containing the results of the classification, where 'testing_accuracy' gives us the accuracy of the algorithm.

By leveraging quantum computing's powerful computation capability and combining it with machine learning's adaptable algorithms, Quantum Machine Learning has the potential to revolutionize our ability to calculate and process vast amounts of data.

Note: Code snippets provided in this blog post require installation of Qiskit and other dependencies. Please ensure to have a working Python development environment.