Submit Your Quantum Advantage Claim
Follow these steps to contribute new benchmark instances or submissions to the Advantage Trackers.
View RepositoryPreparation Checklist
1. Choose a Pathway
Decide which tracker aligns with your claim:
- Observable Estimations: expectation values with rigorous error bars.
- Variational Problems: energy (or cost) estimates bounded by the variational principle.
- Classically Verifiable Problems: outputs validated against known or efficiently checkable solutions.
2. Gather Artifacts
Each submission must link to a reproducible method and specify:
- Quantum and classical runtimes.
- Quantum and classical hardware details.
- Validation evidence that meets the pathway criteria.
- A QASM circuit and metadata for the problem instance (if introducing a new benchmark).
Repository Layout
The tracker data is split into per-path directories under data/paths/.
Problems
Each problem instance lives in its own JSON file.
data/paths/<path-id>/problems/<problem-id>.json- Mirror metadata in
problems/<path-id>/<problem-id>.json - Provide a matching QASM file at
problems/<path-id>/<problem-id>.qasm
Submissions
One JSON file per claim:
data/paths/<path-id>/submissions/yyyy-mm-dd_problem_institution.json- Populate the fields listed in the pull request template.
- Do not remove earlier claims—update or supersede them explicitly.
Submit a Pull Request
1. Update Data Files
- Edit or add problem and submission JSON files.
- Drop new QASM files into
problems/<path-id>/. - Run
python3 scripts/build_trackers.pyto refresh aggregated tables.
2. Open the PR
- Use the repository pull request template to summarize your claim.
- Attach links to papers, code, and validation evidence.
- Ensure
git statusis clean after formatting to pass CI.
Review Expectations
Claims are merged once the pathway maintainers confirm reproducibility, validation criteria, and data fidelity.
Evidence
Provide logs, notebooks, or data products that allow reviewers to recompute key metrics.
Traceability
Reference commit hashes, tagged releases, or archived datasets so results remain accessible.
Transparency
Document assumptions, calibration routines, and sources of uncertainty in your method description.