71. Project 6 — Reproduce a Research Paper

The pipeline (Chapter 53 is the method; this is the execution):

paper → mathematical model → implementation → simulation → experiment → comparison → technical report

Specification. Choose one recent (last 3 years) paper with a simulator-reproducible core claim — the sweet spots: a transpiler/routing benchmark, a decoder evaluation, a variational-algorithm study, a resource-estimate improvement. Constraints from Ch. 53.2: code-adjacent (their repo exists for checking, after your attempt), bounded (one figure), verifiable (numbers to match). Build the reproduction as a public repo: spec.md (the Ch. 53.3 extraction — every equation instanced at small n, every parameter listed, UNSPECIFIED markers where silent), implementation from spec (their code consulted only after your version runs, differences logged), the target figure reproduced with fresh seeds and error bars, the comparison (their number vs. yours, within-joint-uncertainty verdict), and the deviation ledger (Ch. 53.7's format — paper's value, yours, why, measured impact).

Milestones. M1: paper chosen, three-pass read done (Ch. 52), spec written. M2: implementation complete from spec; property tests green. M3: target figure reproduced (or the failure investigated to root cause — both are M3-complete). M4: deviation ledger finalized; comparison written with statistics (Ch. 53.6's standards). M5: published — repo finalized, blog post or ReScience-style write-up (Ch. 53.8), email to the authors with your findings.

Acceptance criteria. A stranger could regenerate your figure with one command; every deviation has a measured impact; the verdict (confirmed / refined / could-not-reproduce) is stated with its statistical basis; the write-up is neutral in tone (Ch. 53.7's discipline) and public.

What it proves. Chapter 53's pipeline executed end-to-end — you have done research-cycle work: extraction, independent implementation, statistical comparison, and public, accountable communication. This is the project that converts "knows quantum computing" into "does quantum research." Effort: 3–5 weekends.