Technical notes and complete deposit (code, synthetic data, preregistrations, decision log, independent audits) for three preregistered synthetic studies of drift-aware decoding of a distance-5 repetition code with one drifting data qubit. QP-02-H7, a kill test at equal information (1850 fresh trajectories): a hard keep/exclude switch beat neither equal-information baseline; verdict NO_SEPARATION. The soft reweighting policy P9 had an 8.6% [95% CI 7.1%, 10.0%] lower logical error rate than the switch, and a change-aware in-epoch estimator P10 3.1% [1.4%, 4.9%]. The hard action is not the source of the advantage measured in earlier runs. QP-02-H8, a preregistered 2x2 decomposition of the P9 versus P10 difference (1225 fresh trajectories): both main-effect labels POSITIVE; under the frozen attribution rule the estimator procedure is attributed to P9's advantage and the protocol regime offsets part of it. QP-02-H9 addendum: with a second qubit co-drifting with the first, the procedure contrast favours the structure-free estimator (run 002, labels NEGATIVE; an outcome-aware validation after run 001, which was INVALID_RUN and is reported beside it, never pooled). All data are simulated; no quantum hardware and no personal data are involved. Code (*.py, *.ps1) is MIT licensed, all other content CC BY 4.0. Start with DEPOSIT_README.md inside the archive.
Technical notes and complete deposit (code, synthetic data, preregistrations, decision log, independent audits) for three preregistered synthetic studies of drift-aware decoding of a distance-5 repetition code with one drifting data qubit. QP-02-H7, a kill test at equal information (1850 fresh trajectories): a hard keep/exclude switch beat neither equal-information baseline; verdict NO_SEPARATION. The soft reweighting policy P9 had an 8.6% [95% CI 7.1%, 10.0%] lower logical error rate than the switch, and a change-aware in-epoch estimator P10 3.1% [1.4%, 4.9%]. The hard action is not the source of the advantage measured in earlier runs. QP-02-H8, a preregistered 2x2 decomposition of the P9 versus P10 difference (1225 fresh trajectories): both main-effect labels POSITIVE; under the frozen attribution rule the estimator procedure is attributed to P9's advantage and the protocol regime offsets part of it. QP-02-H9 addendum: with a second qubit co-drifting with the first, the procedure contrast favours the structure-free estimator (run 002, labels NEGATIVE; an outcome-aware validation after run 001, which was INVALID_RUN and is reported beside it, never pooled). All data are simulated; no quantum hardware and no personal data are involved. Code (*.py, *.ps1) is MIT licensed, all other content CC BY 4.0. Start with DEPOSIT_README.md inside the archive.
Amendment 5 (2026-10-09), deposited before any S2 data. The registered question, estimand, test, seeds (1000-1079), shots (2,000), backend and qubit rules and 600 s ceiling are unchanged. What changes is how sampler calls become IBM jobs: one job per call (960 jobs, about 2,900 billed seconds) cannot fit the ceiling because of per-job billing, so seeds run in two blocks of 40 and every round of a block is one multi-pub job (12 jobs, about 540 s expected). Both arms of a seed share each job; the resulting shared-job exposure is accepted, with the primary test unchanged and per-block estimates reported as a secondary diagnostic. There is no statistical stopping and no interim analysis. Author decisions (shots, blocks, exposure) were made on 2026-10-09. Files: S2_AMENDMENT_5.md, AMENDMENT_5_MANIFEST.json (hashes of the implementation snapshot, commit 3332e893, and of the documents) and rqc-s2-amendment5-5fc50b0c.zip (code, documents and tests at commit 5fc50b0c). The original pre-registration and its frozen tooling are included unchanged. Not approved for execution: a run requires a recorded approval bound to the exact run payload. Pre-registration, deposited before any hardware job. On one IBM Heron backend (alphabetical-first rule; lowest-CZ-error 8-qubit path), with depth, two-qubit gate count, shots and session equal, does feedback-driven local search over single-qubit angles raise the fresh-re-measured, circuit-normalized XEB of the selected circuit compared with random search, at equal output concentration? Primary estimand: β0 of d = β0 + β1·Δc (one-sided, α = 0.05); 80 fresh seeds (1000–1079); 10-minute QPU ceiling. Simulation (S1) on an ibm_fez noise model and on Google's Willow noise model disagreed (β0 +0.010 and −0.003, both intervals include 0); the test is powered for effects of about 0.012 or larger and a null does not rule out smaller ones. Code frozen at github.com/osiris-dnalang/osiris-cli commit 86e3a3f4 (archive included).
This preregistration describes a confirmatory study of an upstream conditioning architecture in a quantum error-correction workflow. The project’s purpose is to test whether an intervention applied prior to the primary decoder produces a real and reproducible decoder-relevant effect, remains compatible with more than one downstream decoder family, generates a measurable exposed-observable statistical footprint, and improves or preserves performance across a preregistered regime rather than only at a single operating point. The study is intentionally limited to architecture-level evaluation, with the public claim set confined to exposed observables, frozen transformations, preregistered statistical tests, explicit negative controls, and bounded interpretation rules. The confirmatory design is organized around five preregistered tests: category existence, decoder compatibility, structural footprint, observable statistical signature, and regime robustness. Across these tests, the central comparison is between a baseline control pipeline and a conditioned pipeline in which the upstream intervention is applied while all non-conditioning settings are held fixed. The study uses frozen observables and frozen execution rules, including matched seeds, shot counts, code distances, physical error rates, geometry, and baseline decoder settings, so that any supported effect can be interpreted as evidence about the architecture under test rather than as a byproduct of post hoc tuning. To preserve a clean-room boundary, the registration restricts primary analysis to exposed, decoder-facing quantities only. These include logical failure, an exposed defect-count summary, and an exposed decoder-difficulty proxy, along with a locked scalar error-load construction derived from control-only standardization. The study does not rely on decoder-internal state variables, hidden confidence traces, or implementation-specific internal diagnostics in its primary confirmatory analyses. This structure is intended to let readers evaluate whether the architecture leaves a reproducible footprint in decoder-facing behavior without disclosing reconstruction-level mechanism details. Expected outcomes are defined in confirmatory rather than speculative terms. If the preregistered hypotheses are supported, the study is expected to provide evidence that an upstream conditioning intervention exists as a real architectural category in QEC, acts before the primary decoder, is compatible with downstream decoding, changes decoder-relevant exposed structure, and produces a reproducible observable statistical footprint in logical-failure behavior. If support is absent, mixed, or null, those outcomes will still be recorded under the frozen rules without rescue through post hoc metric changes, reweighting, or reinterpretation. In either case, the expected result of the registration is a bounded, auditable public record of whether this architecture class shows reproducible category-level behavior under a preregistered confirmatory design. The preregistration is not intended to establish a unique mechanism identity, identify a privileged implementation, disclose implementation-specific reconstruction details, or support claims about exact targeting logic, suppression laws, subset-sizing logic, telemetry-governed control, or broader cross-domain theory narratives. Public-facing interpretation is therefore intentionally limited to category validation and exposed-observable behavior. This is meant to support rigorous public evaluation while preserving strict clean-room limits around protected implementation detail.
We report five preregistered experiments on the IQM Garnet 20-qubit square-lattice processor, each with its decision rule committed before any hardware data existed and each reported regardless of outcome. First, the even-over-odd parity-sector leakage asymmetry previously measured on IBM Heron r2 replicates on Garnet at preregistered power (relative asymmetry +0.068, one-sided p = 1.1 × 10−5). Because the noiseless parity leakage of these number-conserving XY-Trotter circuits is exactly zero, this is a statement about device noise. Second, the asymmetry is not stable: over ten execution windows it varies beyond shot noise at fixed depth (Cochran Q = 19.01 on 9 degrees of freedom, p = 0.025), more strongly when windows are pooled by provider calibration set (p = 0.0083), and an isolated sign reversal at depth 10 does not replicate. Third, a confirmatory campaign that takes the calibration epoch as its independent unit — twelve epochs, 688,128 shots — finds the asymmetry larger at depth 8 than at depth 12 (mean contrast +0.0173, one-sided p = 0.026); the lower bound is +0.0030 and three of twelve epochs have the opposite sign. Fourth, a calibration-aware layout optimiser developed on heavy-hex devices fails its frozen transfer rule: it is better than automatic placement at chain lengths 8 and 16 and worse at 12, and a powered follow-up confirms the reversal. All preregistrations, calibration snapshots, job identifiers, raw counts and analysis scripts are public. All quantum-hardware results were obtained on IQM Garnet through IQM Resonance under an extended-access research grant. Preregistrations, raw counts and analysis code: https://github.com/anulum/scpn-quantum-control
UIP is a candidate framework formalizing coherence, drift, inheritance, resonance, waste, and non-harm constraints as measurable control-theoretic and information-theoretic objects, governed by dC/dt = G(R,U) − L(D,W) subject to a hard non-harm envelope. This Phase 1 record contains three artifacts: (1) the preprint; (2) Appendix A — a single noisy qubit memory in which every UIP object is numerically specified, six preregistered falsification tests passed, and the fixed-point condition G = L predicted steady-state drift D* = 0.1032 vs 0.1051 simulated (within 2%); (3) Appendix B — a two-qubit test of inheritance-guided correction allocation, which FAILED its preregistered falsification condition in both version 1 and a preregistered version 2 across a full coupling sweep, reported in full with the diagnosed mechanism and a sharpened restricted-access boundary hypothesis (v3) left open for independent testing. All results are consistent with standard quantum mechanics, quantum information theory, and feedback control. This work demonstrates reduction to known science (Criterion 1) only; novel prediction (Criterion 2) and independent replication (Criterion 3) remain open. UIP is a candidate framework, not established science. Refutations are solicited as explicitly as confirmations; simulation code is published openly for reviewers to reproduce, vary, and attempt to break every result. Simulation code (open source): https://github.com/derekhone/uip-phase1-testbeds Version 2 adds Appendix C: a preregistered v3 test of inheritance-guided control under restricted access. Predictions were locked before execution. Result: robust pass at weak coupling (+3.5% over best baseline, N=100 seeds); loses at strong coupling. Full boundaries reported. Code: https://github.com/derekhone/uip-phase1-testbeds Version 3 completes Phase 1. Adds Appendices D–G and the synthesis manuscript. D/E: preregistered crossover point and zero-free-parameter law s*(γ) = arcsin(√(γ(1−C̄*)/C̄*)), C̄* = 1/1.45, confirmed across a γ-sweep (r = 0.976), with three-qubit generalization. F: negative result — decision quantities fall 12.19% (v3) and 22.06% (v4) short of a beam-search near-optimal adversary; control-sufficiency claim falsified and closed. G: negative result — C̄* does not transfer to a structure-matched classical system (no crossover exists); universality claim retired. Synthesis manuscript: four confirmed preregistered predictions, three falsified claims. No new physics is claimed. Version 4 adds Appendix H — the first hardware test, run on IBM Quantum (ibm_kingston, Heron r2, job d99brisqp3as739tudkg). The Appendix A fixed-point law, with the per-cycle decay computed from published calibration T2, FAILED its preregistered ±10% kill-condition (predicted D* = 0.0302, measured 0.1463, deviation 383.7%). Diagnosis: bare-idle decay is governed by T2*, not echo-T2; the fixed-point structure (flat steady state) was observed, but the input noise model is falsified on hardware. Sharpened hypothesis H2 (in-situ rate measurement or echo insertion) is preregistrable and left open. Preregistration was committed to the GitHub repository before execution. Ledger: 4 confirmed / 4 falsified — including one rejection delivered by hardware outside our control. FINAL VERSION — PHASE 1 CLOSED (July 12, 2026). This version adds Appendix I (H2a echo-protected hardware test) and the final complete synthesis paper, closing the Phase 1 falsification program. Final ledger: 4 claims confirmed, 5 falsified — including two preregistered hardware rejections of the author's own predictions on IBM Quantum's ibm_kingston processor (Heron r2, 156 qubits), both published unedited the same day. CONFIRMED (simulation): (A) single-qubit recirculation fixed point, steady-state drift predicted analytically within 2%; (C) restricted-access regime where inheritance-guided control is advantageous; (D–E) quantitative crossover law s*(γ), confirmed as preregistered. FALSIFIED: (B) inheritance-guided two-qubit control vs. static baseline, two preregistered versions; (F) near-optimality claim — 12.19% gap vs. a 5% threshold; (G) the universality claim — the word "Unified" was formally retired as a consequence; (H) the fixed-point law with published-calibration inputs on hardware — 383.7% deviation vs. a preregistered ±10% window (job d99brisqp3as739tudkg); (I) the same law with echo-protected idles correcting the diagnosed T2/T2* input error — 81.9% deviation (job Post-closure replication note (2026-07-13): The fixed-point law falsified in Appendix I was re-tested on a second device (ibm_marrakesh, qubit 2, preregistered, job d9airde6hjac73fegau0). Result: FAIL — measured D* = 0.0529 vs. predicted 0.0407 (30.1% deviation). The falsification is device-independent (kingston 81.9%, marrakesh 30.1%). Ledger unchanged: 4 confirmed / 5 falsified.
Proposal to the BlueQubit Quantum Flywheel open call (2026). Track 2 (breaking quantum advantage claims) with a Track-1 hardware component and a Track-3-adjacent seed. Aim 1 — execute the author's own hash-verified K₈ τ-sweep pre-registration (10.5281/zenodo.17918774, never run as written) exactly as specified on three IBM Heron backends under one pre-declared backend amendment; publish the outcome either way. Aim 2 — adversarial classical re-analysis of the author's 156-qubit / 10⁶-shot and 1,430-circuit "entropy suppression" datasets (10.5281/zenodo.19656600, 10.5281/zenodo.19505907) with coverage-adjusted estimators and tensor-network reproduction. Aim 3 — pre-registered hardware test of whether an auditable rule-evolving controller (organism_sim / bridge, 10.5281/zenodo.22862567) helps evolutionary dynamical-decoupling search under real calibration drift; four arms, binding criterion stated in PREREGISTRATION.md before any job. Includes PREREGISTRATION.md (binding criteria committed before submission of any hardware job), prepared form answers, and the supporting audits of the author's 60 prior records (HONEST_ASSESSMENT, FLYWHEEL_ASSESSMENT, PUBLICATIONS index). The archive is git archive of tag v1.0.0.
Six frozen interference predictions, independent calibration and a count-preserving memory test g∗=(1,c∗,2c∗2−1,c∗2,1,c∗2),c∗=cos8π. Can a retained reference recover interference that two fresh references cannot? This prospective protocol freezes the signed coherence for six histories: null, single, shared forward, fresh forward, shared echo and fresh echo. The forward and echo memory contrasts are respectively minus and plus sin2(π/8)=0.1464466094…. The paper derives the complete reference-partition map, specifies 24 history/phase settings, preserves every raw attempt including loss, and separates source coefficients from independent apparatus calibration. It fixes a disjoint training sample, 2.4 million validation attempts, simultaneous finite-sample rejection and equivalence criteria, and held-out scores against explicitly declared comparators. A second, independently timestamped laboratory capsule is required before validation data are acquired. What the test establishes. A qualified joint pass validates this QTT optical construction and its reference-memory pattern in the specified apparatus. The matched controlled-unitary quantum-mechanical model predicts the same pattern with the same calibration and no additional fitted parameters. Therefore this benchmark does not distinguish QTT from matched QM or independently establish an absolute clock. Any predictive advantage over a fitted memory-free reduction or history-specific harmonic fit is reported separately. Design-stage preregistration. The deposit includes the full paper, frozen machine-readable predictions, analysis code, independent algebra and mutation checks, synthetic operating checks, and exact reconstruction checksums. No new laboratory result is reported. The earlier 95.76% injected-template retention motivates count-preserving processing; it is neither a natural-signal detection nor an experimental correction factor. QTT source dependency: Balanced-Rail Optical Intertwining Theorem. The necessary mathematics is reproduced here. Related physical reference qualification: Physical-Terminal Reference Framework. The current PDF is the default preview; the ZIP contains the complete reconstruction package.