Cirq: Google’s Open-Source Python Quantum Circuit Framework
Python programming framework Cirq is open-source. With them, you may design, edit, modify, optimise, and activate quantum circuits. Google AI Quantum released Cirq as a public alpha on July 18, 2018. This Apache 2 license allows it to be modified or integrated into any open-source or commercial program.
Designed for Noisy Intermediate Scale Quantum computers and algorithms. NISQ computers are noise-sensitive systems with 50–100 qubits and high-fidelity quantum gates, or fewer than a few hundred. Current quantum algorithm research focusses on NISQ circuits, which work without error correction, because they may offer a quantum advantage on near-term devices.
The methodology helps researchers determine whether NISQ quantum computers can tackle real-world computational problems by focussing on short-term challenges. It provides abstractions for NISQ computers, where hardware details are crucial to cutting-edge results.
It lets users organise gates on the device, plan time within quantum hardware constraints, and specify gate behaviour using native gates to precisely manage quantum circuits. Cirq data structures are optimised for building NISQ quantum circuits to help users maximise their use. Cirq's architecture targets NISQ circuits, so researchers and developers can experiment with them and find noise-reducing ways.
Cirq has many quantum circuit capabilities:
Circuit Manufacturing/Manipulation This provides a flexible and easy interface for designing quantum circuits. User-defined measurements, qubits, and quantum gates can match their mathematical description. Custom and flexible gate definitions are supported by the framework. Users can learn about moments, insertion strategies, and qubit gates to build quantum circuits. Symbolic variable parameterised circuits work too. Circuits can be chopped, sliced, and diced to improve them.
Hardware modelling Hardware restrictions drastically affect a circuit's feasibility on modern hardware. Cirq devices can manage these constraints. Noise and hardware device modelling are supported.
Modelling Cirq includes density matrix and wave function integrated quantum circuit simulators. These simulators can handle noisy quantum channels with full density matrix or Monte Carlo simulations. Cirq connects with cutting-edge wave function simulators like qsim for high-performance simulation. The Quantum Virtual Machine (QVM) can emulate quantum hardware with these simulators. Quantum circuit simulation is vital for algorithm development and testing before hardware implementation.
Optimising and Transforming Circuits Optimisation, compilation, and circuit transformation are framework functions. Many quantum circuit optimisers for hardware testing are included.
Implementing Hardware Cirq is cloud-integrable with larger simulators or future quantum hardware. It simplifies quantum circuit operation for quantum processors. Cirq lets users run quantum circuits on Google's quantum hardware via its Quantum Computing Service. The Google AI Quantum team designs circuits for Google's Bristlecone processor using Cirq as a programming interface and plans to make it cloud-available. It can also submit quantum circuits to cloud platforms like Azure Quantum for execution on QPUs, IonQ, and Quantinuum simulators.
Interoperability It supports SciPy and NumPy. Cross-platform, it supports Linux, MacOS, Windows, and Google Colab.
Applications and use cases
Cirq is crucial to Quantum Machine Learning (QML), which combines machine learning and quantum computing. QML uses quantum computing to speed up machine learning. Cirq implements quantum machine learning algorithms on NISQ devices. Quantum neural networks mimic neuronal behaviour to find patterns and predict.
Quantum Support Vector Machine (QSVM) algorithms can be built on Cirq using quantum circuits for linear algebra operations and quantum feature maps to encode data into quantum states. Quantum SVM algorithms exist. Cirq Quantum implementations of popular machine learning methods have been made.
Cirq can be used to test quantum algorithms like the Quantum Approximate Optimisation Algorithm (QAOA) and Variational Quantum Eigensolver (VQE) for machine learning applications. QAOA, a hybrid quantum-classical technique, handled combinatorial optimisation problems.
Quantum computers, which can be duplicated using Cirq, are perfect for optimisation problems because they can investigate numerous answers. This reduces training time and finds optimal machine learning model parameters.
Cirq is also used in machine learning to create quantum versions of classical algorithms like decision trees, apply quantum error correction to improve model robustness, simulate chemical reactions for drug discovery, and encode classical data into quantum states. Many quantum gate functions are needed to build complex machine learning circuits with Cirq.
This is used for QML and Google quantum processor end-to-end tests. Early adopters of Cirq simulated quantum autoencoders, implemented QAOA, integrated it into hardware assessment software, simulated physical models like the Anderson Model, and integrated it with quantum compilers.
Close-term algorithms in quantum chemistry challenges can be facilitated by OpenFermion-Cirq.
Over 200 people contributed. Researchers, software engineers, technical writers, and students are encouraged to contribute to its open and inclusive community. Communities organise weekly virtual open source meetings including Quantum Circuit Simulation Weekly Sync, TensorFlow Quantum Weekly Sync, OpenFermion Weekly Sync, and Cirq Weekly Sync.
On Quantum Computing Stack Exchange, utilise the cirq tag for Cirq questions. GitHub provides good initial issues and contribution criteria for programming contributors. Larger features use an RFC (Request for Comment). The Quantum AI website's Cirq home page offers examples, reference documentation, and text-based, Jupyter notebook, and video tutorials. Releases occur every three months.
Google Quantum AI's open-source software stack includes Cirq. It includes the open-source hybrid quantum-classical machine learning library TensorFlow Quantum. It also interacts with OpenFermion family libraries for chemistry and material science. More integrated Google stack products include Qualtran for fault-tolerant quantum computing, Stim for huge Clifford circuits and quantum error correction, ReCirq for Cirq experiments, and Qsim for high-performance simulation. Use Cirq to submit quantum circuits to third-party cloud services like Microsoft Azure Quantum to access IonQ and Quantinuum hardware.
Finally, Google's Quantum AI team developed Cirq, an open-source Python framework for NISQ quantum computer creation and experimentation. Its full circuit design control, robust simulation capabilities, and interfaces for executing circuits on real quantum hardware and integrated simulators make it crucial in quantum machine learning research.
