Skip to main contentIBM Quantum Documentation Mirror

Overview of noise management techniques

When executing quantum workloads, there are multiple ways to reduce the impact of noise. The open-source Qiskit addons provide error mitigation and suppression techniques that integrate directly into your development workflow, while Qiskit Runtime applies advanced error mitigation strategies automatically when jobs are submitted for execution. This page indexes all available tools and features across both options to help you choose the right approach to manage noise when building quantum workloads.


General noise management techniques

  • Directed execution model

    Fine-tune error mitigation and other techniques by capturing design intents on the client side and shifting, and shifting the costly generation of circuit variants to the server side.

  • Dynamical decoupling

    Inserts pulse sequences on idling qubits to suppress coherence errors caused by unwanted interactions between qubits during circuit execution.

  • Pauli twirling

    Noise-tailoring technique that transforms any noise channel into a Pauli channel that has a more specific structure; often combined with other error mitigation techniques that work well with Pauli noise.

  • AQC-Tensor Qiskit addon

    Users can compile the initial portion of a circuit into a nearly equivalent approximation of that circuit, but with fewer layers.


Error mitigation for expectation value estimation

  • Twirled readout error extinction (TREX)

    Error mitigation tool within Qiskit Runtime which mitigate the effects of measurement errors by randomly substituting them with a twirled measurement sequence.

  • Zero-noise extrapolation (ZNE)

    Error mitigation technique that computes the expectation value at different noise levels, and then estimates the ideal result by extrapolating the noisy expectation value results to the zero-noise limit.

  • Probabilistic error amplification (PEA)

    ZNE technique that involves running preliminary experiments to learn a twirled noise model of the circuit, then uses this model to perform a more accurate error amplification.

  • Probabilistic error cancellation (PEC)

    Returns an unbiased estimate of the expectation value, at the expense of greater overhead than other techniques such as ZNE. It extrapolates output of the ideal circuit by executing different noisy circuit instances.

  • PEC with shaded lightcones

    A modified PEC technique that uses Pauli propagation to reduce the number of error terms accounted for in a noise model according to the specifics of the target observable.

  • Operator backpropagation (OBP)

    Uses a method based on Clifford perturbation theory to reduce circuit depth by trimming operations from its end at the cost of more operator measurements.

  • Propagated noise absorption (PNA)

    Technique for mitigating errors in observable expectation values by 'absorbing' the inverses of learned noise channels into an observable using Pauli propagation.


Error mitigation for sampling results

  • Sample-based quantum diagonalization (SQD)

    Implements a technique for finding eigenvalues and eigenvectors of quantum operators, such as a quantum system Hamiltonian, using quantum and distributed classical computing together.

  • SQD for HPC

    An HPC-ready implementation of the SQD addon. It is written in modern C++17 standards and is designed to create a single compiled binary for use with MPI.

  • Matrix-free Measurement Mitigation

    Matrix-free Measurement Mitigation (M3) is a package for scalable quantum measurement error mitigation that can be computed in parallel.