Skip to content
Nathan Shammah edited this page Feb 11, 2022 · 32 revisions

Welcome to the mitiq wiki!

Mitiq Roadmap (summary, look here for 2022 planning)

Error mitigation techniques:

  • Current:

    • Zero-Noise Extrapolation (ZNE)
    • Probabilistic Error Cancellation (PEC)
    • Clifford Data Regression
    • Readout error mitigation
  • Planned:

    • Dynamical Decoupling
    • Noise-Extended Probabilistic Error Cancellation (NEPEC)
  • Candidates for the future (not planned):

    • Symmetry-based techniques, such as: Error Suppression by Derangement / Virtual Distillation, Symmetry Verification.

Integrations / APIs

  • Current:

    • Cirq
    • Qiskit
    • PyQuil
    • Braket
    • PennyLane
  • Future:

    • QuTiP
    • OpenPulse.

Mitiq 2022 Roadmap

Mission: Make error mitigation easily available to everyone running programs on quantum computers. Objective: Mitiq should be so effective that every quantum program should compile through mitiq.

Q1

  • Scope and prototype Dynamical Decoupling (DD): Literature Review. Notebooks. Write a function which inserts spin echoes on idle qubits in a Cirq circuit that shows some possible improvement in estimated expectation values.
  • Benchmark project: Run 3 error mitigation techniques (e.g., ZNE, PEC, CDR) on 2 tasks (e.g., RB, mirror circuits) on 3 different hardware devices (e.g., multiple superconducting qubits, ion chips).
  • Mitiq Maintenance and enhancements: Identify and fix at least 6 open issues with a high-priority label (p0 or p1).

Q2

  • Implement DD and make public release: Incorporate tools into a Mitiq module/technique: (i) Use conversions for frontend IO, (ii) define an execute_with_dd function with the canonical Mitiq signature. Continue Benchmark project: Finalize results and data analysis
  • Mitiq Maintenance and enhancements: Identify and fix at least 6 open issues with a high-priority label (p0 or p1).
  • Prototype QIR support: and propose full implementation timeline

Q2

  • Test DD on hardware: Use Mitiq’s DD on a hardware provider with a benchmark circuit to demonstrate improved performance (more accurate results) in the docs.
  • Benchmarks: Write and upload paper with results
  • Prototype readout error mitigation (REM)
  • Mitiq Maintenance and enhancements: Identify and fix at least 6 open issues with a high-priority label (p0 or p1).

Q4

  • Implement REM and test on hardware: Write a function to compute readout errors (confusion matrices) on a set of qubits on a given backend.
  • Mitiq Maintenance and enhancements: Identify and fix at least 6 open issues with a high-priority label (p0 or p1).

Talks and Events

Clone this wiki locally