This repository contains Python scripts that demonstrate quantum image classification using the Cirq and PennyLane libraries with the MNIST dataset. The scripts create quantum circuits and apply variational classifiers to predict the digit represented in a given test image.
Guala, D., Zhang, S., Cruz, E. et al. Practical overview of image classification with tensor-network quantum circuits. Sci Rep 13, 4427 (2023). https://doi.org/10.1038/s41598-023-30258-y
Ménard, A., Ostojic, I., Patel, M., & Volz, D. (2020). A game plan for quantum computing. McKinsey Quarterly. https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/a-game-plan-for-quantum-computing
To run the scripts, you need to have the following prerequisites installed:
- Python (version 3.6 or higher)
- Cirq library (for the Cirq script)
- PennyLane library (for the PennyLane script)
- TensorFlow library (for the Cirq script)
- NumPy library
- Matplotlib library (for visualizing the PennyLane circuit)
-
Clone this repository:
git clone https://github.com/your-username/quantum-image-classification.git
-
Install the required libraries:
pip install cirq pennylane tensorflow numpy matplotlib
-
Navigate to the cloned repository:
cd quantum-image-classification -
Run the Cirq script:
python cirq_image_classification.py
The script will load the MNIST dataset, create a quantum circuit using Cirq, and build a classical model using TensorFlow's Keras API. It will then train the model and print the circuit diagram.
-
Navigate to the cloned repository:
cd quantum-image-classification -
Run the PennyLane script:
python pennylane_image_classification.py
The script will load the MNIST dataset, create a quantum circuit using PennyLane, and apply a variational classifier to predict the digit represented in a test image. The predicted digit will be printed in the console, and the circuit diagram will be saved as
pennylane_circuit.png.
The Cirq script creates a quantum circuit using Cirq's GridQubit and applies quantum operations to the qubits. The circuit diagram is obtained using the to_text_diagram() method.
The PennyLane script defines a quantum circuit using PennyLane's qnode and encodes the input image into the quantum circuit using RX gates. The StronglyEntanglingLayers template is then applied to the circuit, followed by measurements of the expectation values of the Pauli-Z operator on each qubit.
The Cirq script builds a classical model using TensorFlow's Keras API. The model consists of a flatten layer, dense layers with ReLU and softmax activations. The model is compiled and trained using the GPU.
The PennyLane script defines a variational classifier in the variational_classifier function. It takes the weights and bias as input and returns a classifier function. The classifier function applies the quantum circuit to the input image and computes the dot product of the circuit outputs with the weights, adding the bias term.
Both scripts predict the digit represented in a test image from the MNIST dataset. The Cirq script prints the circuit diagram, while the PennyLane script prints the predicted digit and saves the circuit diagram as pennylane_circuit.png.
The following table compares various quantum computing platforms and their capabilities:
| Quantum Company | Quantum AI Platform | Quantum AI Language | Quantum AI Summary | Quantum Computer | Public Access | Simulated or Actual | Classical-Quantum Hybrid Functionality | Dependencies | Total Qbits | Usage Pros | Usage Cons | Quantum Hardware Types |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Amazon Braket | PennyLane | Python | A cross-platform Python library for quantum machine learning, automatic differentiation, and optimization of hybrid quantum-classical computations. | Various (IonQ, Rigetti, D-Wave) | Yes | Both | Yes | Python 3.7+, PennyLane 0.28.0+, NumPy 1.20.0+ | [11, 5000+] | Integration with AWS, access to multiple quantum hardware providers, seamless integration with classical machine learning frameworks | Requires an AWS account and credits, limited to supported quantum hardware providers | Superconducting qubits, Trapped ions, Quantum annealers |
| IBM Quantum | Qiskit | Python | An open-source quantum computing framework for programming quantum computers, with a focus on quantum circuits. | IBM Quantum Experience | Yes | Both | Yes | Python 3.7+, Qiskit 0.37.0+ | 127 | Large community, extensive documentation, access to real quantum hardware, integration with classical Python libraries | Limited quantum hardware availability, requires knowledge of quantum circuits and algorithms | Superconducting qubits |
| Cirq | Python | An open-source framework for writing, manipulating, and optimizing quantum circuits and running them on quantum computers and simulators. | Google Sycamore | No | Both | Yes | Python 3.7+, Cirq 1.1.0+ | 53 | Intuitive and user-friendly API, extensive documentation and tutorials, supports both simulation and execution on real quantum hardware | Limited access to Google's quantum hardware, primarily focused on gate-based quantum computing | Superconducting qubits | |
| Microsoft Azure Quantum | Q# | Q# (C# integration) | A domain-specific programming language used for expressing quantum algorithms and integrating them with classical code in the .NET environment. | Various (IonQ, Honeywell, QCI) | Yes | Both | Yes | N/A | [11, 32] | Seamless integration with the .NET ecosystem, access to multiple quantum hardware providers through Azure, extensive documentation and tutorials | Requires familiarity with the .NET framework and C#, limited community compared to Python-based frameworks | Superconducting qubits, Trapped ions |
This comparison table provides additional information about the quantum platforms, including their associated quantum programming languages, quantum computers, public access availability, simulation and actual hardware capabilities, classical-quantum hybrid functionality, dependencies, total qubits, usage pros and cons, and supported quantum hardware types.
Contributions to this project are welcome. If you find any issues or have suggestions for improvements, please open an issue or submit a pull request.
Copyright 2024 Eric Yocam
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.
Disclaimer This repository is intended for educational and research purposes.