A Peer-Review Checklist for Quantum Circuits and Experiments

Quantum programs are unusually hard to check alone. Results are probabilistic, hardware is noisy, and a circuit that looks right on paper can encode a different operation from the one intended. The person who wrote it carries a mental picture of what each gate should do, and that picture overrides what the diagram actually says. This checklist is designed for a pair: one author, one reviewer, working through the same items before any circuit or experiment is run on real hardware or reported to colleagues.

How to use this checklist

Print or copy it. The author completes the first column of thinking; the reviewer reads the circuit, notebook or lab notes without being told the intended outcome first, then compares. Record every discrepancy. Do not argue during the first pass. Disagreements are the useful output, because each marks a place where one reader’s mental model differs from the page.

Section A: Problem statement

  • Is the goal written in one sentence a non-specialist colleague could repeat back?
  • Is the expected output stated before the run, including what counts as success?
  • Are the input state and any assumptions about it written down, not remembered?
  • Has the reviewer been given the goal only, so that they can describe the circuit independently?

Section B: Circuit and code

  • Does the qubit ordering used in the code match the ordering assumed in the diagram? Conventions differ between toolkits, and this is a classic source of silent error.
  • Are all gates, angles and units as intended, including whether rotations are expressed in radians or fractions of a full turn?
  • Are measurements placed where intended, and are classical bits mapped to the right qubits?
  • Has any leftover test code, such as a forced initial state or a skipped step, been removed?
  • Can the reviewer reproduce the circuit from a plain description without reading the code first?

For transformations such as the quantum Fourier transform, ordering and phase conventions deserve extra attention. Readers who want the background can consult the explanation of how the quantum Fourier transform drives many algorithms, then examine the reviewed circuit for the typical bit-reversal and sign choices.

Section C: Simulation before hardware

  • Was the circuit run on a noiseless simulator, and does the output match the expected distribution?
  • Is there a small case with a known answer that the circuit reproduces?
  • Were random seeds fixed and recorded?
  • Has the reviewer rerun the simulation independently, on their own machine or environment?

The last item is easy to skip and valuable to keep. A result that only the author can reproduce is not yet a result.

Section D: Noise and hardware assumptions

  • Which device or noise model is assumed, and when was its calibration data taken?
  • Is the number of shots stated, and is it enough for the precision claimed?
  • Are error-mitigation steps listed, with the reason for each?
  • Is the circuit depth reasonable for the hardware’s coherence limits, as stated by the provider?

Section E: Interpretation and claims

  • Do the claims in the write-up match what the data supports, no more?
  • Are comparisons with classical methods fair, using a sensible baseline?
  • Are uncertainty and variation reported alongside single numbers?
  • Has the reviewer tried to write a sentence that contradicts the claim? If they cannot, that is encouraging; if they can, investigate.

Applications deserve similar scrutiny. Optimisation proposals, for instance, often compare a quantum approach with a weak classical baseline. The overview of the way quantum methods approach tough optimisation tasks gives the context needed to judge whether a comparison is fair. Likewise, claims in machine learning deserve a second reading; see the article on what quantum techniques may mean for learning algorithms for the surrounding picture.

Section F: Records and handoff of findings

  • Is every discrepancy found in review logged, with how it was resolved?
  • Does the notebook record library versions, backend names and dates?
  • Are the reviewer’s name and the date of review noted in the project log?
  • Will the author review the reviewer’s next piece within an agreed window?

A worked example of one discrepancy

Suppose an author builds a three-qubit circuit meant to prepare a particular entangled state and then measure all qubits. In the diagram, the top wire is the most significant bit. In the code, the toolkit treats the first qubit as the least significant. Both are legitimate conventions. The author, knowing the intent, reads the histogram of results and sees what looks like the right pattern, because the labels in their head match the diagram. The reviewer, told only the goal, reconstructs the circuit from the code and describes a state with the bits reversed. The mismatch takes five minutes to find and would have cost a poster, a talk or a paper revision had it survived. This is the kind of discrepancy section B is built to expose, and it is a reminder that checks on the mapping between notation and code belong on every list.

Scoring the audit

You do not need a numeric score. Use three outcomes per item: confirmed, discrepancy or not applicable. A sheet with zero discrepancies on first pass deserves suspicion; either the circuit is simple or the review was shallow. Aim for specific, recorded observations over reassurance. A short log such as “qubit order differs between diagram and code” is worth more than a general approval.

Common failure modes in quantum reviews

  1. Reviewer told the answer first. They then confirm instead of examine. Provide the goal, not the expected result.
  2. Review only the code. Diagrams, notes and result plots hide as many errors as the script.
  3. Trusting simulator output as proof. Simulators confirm logic, not hardware behaviour.
  4. Too senior a reviewer only. Junior colleagues ask the questions experts have stopped asking.
  5. One-way arrangements. If only one person reviews, the habit fades. Make it mutual.

Paperwork outside the lab

Researchers also fill in grant forms, travel authorisations and conference registrations, where one wrong detail triggers an administrative delay. The pairing habit helps there too. Even unrelated websites show why: KneeHospitals, a site aimed at people looking at care abroad, whose hospital and doctor pages are split by nation and which includes enquiry forms for prices and visits, asks visitors to type personal information accurately. We offer no view on any medical choice; the practical lesson is that a friend should read form entries aloud before they are sent.

Adapting the list to your group

Teams working on different problems will weight sections differently. A group building variational circuits might expand section D on noise and shot counts; a group writing educational notebooks might focus on section B and the clarity of explanations. Keep the structure but swap items freely, and prune any that never produce a finding after several rounds. A checklist that grows without limit stops being used.

Next steps

Choose one circuit from your current project. Hand a reviewer only the goal and the notebook. Work through sections A to F in a single sitting of about an hour, and log every discrepancy. Then swap roles on their next piece. After three rounds, review the log for recurring items and add them to the checklist. That customised list will serve your group better than any generic template, including this one.

Related Posts

Small Experiments in Quantum Computing: Your Questions Answered

How do you learn quantum computing by doing? Answers to common questions about tiny simulator experiments, from changing one gate to reading results.

Read More

How One Quantum Team Escaped Its Folder of Final_v2 Circuits

A fictional team story shows how a five-slot naming template turned a cluttered quantum experiment repository into one anyone could read at a glance.

Read More

Quantum Computing on One Page: Where Things Stand

A one-page view of quantum computing today: hardware, noise, algorithms, security and what to do next, plus a calm way to keep your page current.

Read More

A Weekly Review for Learning Quantum Computing, Explained

Quantum ideas build on each other, so shaky foundations show up late. A plain-language look at how a weekly review keeps your learning cumulative.

Read More

6 FinOps Myths Emerging-Tech Teams Believe, and the Facts

Six common myths about FinOps held by teams running experimental workloads, with the facts and a practical fix for each.

Read More

PeopleOps Primer for Technologists, With a Mini-Glossary

A friendly beginner primer on running the employee lifecycle as a measured operation, with a twelve-term glossary for readers who know technology but not HR.

Read More
Subscribe
Notify of
guest
0 Comments
Oldest
Newest Most Voted
0
Would love your thoughts, please comment.x
()
x