Submission-level experiment protocol

A computational application of the theorems should retain:

1. the supplied source snapshot and immutable baseline artifact;

2. data snapshots, labels, architecture, numerical precision, software versions, hardware, and determinism settings;

3. complete phase state, masks, activation, damping, and optimizer-reset policy;

4. named training-order, scoring, regrowth, input-noise, and hidden-noise streams;

5. layerwise integer deletion and regrowth counts, with a declared tie rule;

6. the complete intervention budget vector and separate training cost ledger;

7. all candidate configurations, including failed and unmatched ones, rather than only favorable matches;

8. acute, challenged, and modulation-removed checkpoints, with separate displacement and density labels;

9. an independent validation or simultaneous-selection procedure;

10. absolute post-challenge outcomes and paired degradation estimates;

11. repeated training seeds if claims concern training-procedure averages;

12. margins, operator norms, and implementation errors required by the claimed certificate.

A quantum application additionally requires:

1. a specified encoding and preparation class;

2. Hamiltonian and dissipator normalization, time interval, and accessible control class;

3. complete instruments and failure branches;

4. referenced channel or matrix-unit implementation certificates;

5. calibrated nuisance alternatives;

6. an explicit task and resource model before any advantage claim.

No complete biological correspondence dossier is supplied by the analytic construction.