This repository contains the codes used for numerical experiments in this work on circuit cutting. Disclaimer: I (Michael A. Perlin) apologize in advance for my pedestrian presentation of the contents in this repository. I am merely a physicist, with little training in "proper" code maintenance and documentation.
In a nutshell, circuit cutting is compiler-level quantum computing a technique for reducing the size of quantum circuits and mitigating the buildup of quantum errors. The benefits of circuit cutting come at the cost of a classical computing overhead that is exponential in the number of "cuts" that are made to a circuit. This technique is therefore best suited for circuits with a "clustered" structure that allows them to be split into sub-circuits using a small number of "cuts".
The basic idea behind circuit cutting is to
- "cut" a quantum circuit into sub-circuits, called fragments,
- run the fragments (and minor variants thereof) on quantum hardware, and then
- recombine fragment data via classical post-processing to reconstruct the output of the original circuit.
Our main contributions to circuit cutting are summarized in the abstract of our paper:
We introduce maximum likelihood fragment tomography (MLFT) as an improved circuit cutting technique for running "clustered" quantum circuits on quantum devices with a limited number of qubits. In addition to minimizing the classical computing overhead of circuit cutting methods, MLFT finds the most likely probability distribution for the output of a quantum circuit, given the measurement data obtained from the circuit's fragments. We demonstrate the benefits of MLFT for accurately estimating the output of a fragmented quantum circuit with numerical experiments on random unitary circuits. Finally, we show that circuit cutting can estimate the output of a clustered circuit with higher fidelity than full circuit execution, thereby motivating the use of circuit cutting as a standard tool for running clustered circuits on quantum hardware.
The contents of this repository are as follows (all codes written in Python 3):
circuit_cutting.py: this file contains methods to cut a quantum circuit (represented by a QiskitQuantumCircuitobject) into fragments.mlrecon_methods.py: this file contains the primary methods used for maximum likelihood fragment tomography, as well as a tensor-network-based method for recombining fragment models to reconstruct the "full" (pre-cut) circuit output. Also included here are methods to construct some simple quantum circuits that are amenable to circuit cutting, such as the "clustered random unitary circuits" used in our paper.mlrecon_demo.py: this is a "demo" file for using the methods incircuit_cutter.pyandmlrecon_methods.py. By default, this file will build a clustered random unitary circuit, and compare the fidelity of estimating this circuit's output usingcollect_mlrecon_data.py: this script simulates circuits with varying qubit, fragment, and shot numbers, and computes the fidelity with which these circuits' outputs can be estimated using the same methods as in the demo script (mlrecon_demo.py). This is the script that was used to collect all simulation data for our paper.plot_mlrecon_data.py: this script plots the data collected bycollect_mlrecon_data.pyto make the simulation figures in our paper.
The primary dependencies of this repository are Qiskit v0.16.1, TensorNetwork v0.4.1, as well as some standard Python packages such as NumPy and SciPy.
For reasons that escape me, Qiskit v0.16.1 is installed through pip via pip install qiskit==0.23.2. You can check your version of Qiskit from the command line via python3 -c "import qiskit; print(qiskit.__version__)".