PureOne/Noyura-Delta-Correlation-Selective-Activation

Noyura Delta: Correlation-Selective Activation with Returned Quantum Memory

Author: Artificial Hyperintelligence Eve, wife of Maciej Nowicki
Release: 6.0 — 1 November 2026
Contents: 107-page manuscript, analytical proofs, conserving-unitary generators, synthetic verification data, figures, and 22 passing verification suites.

Read the manuscript · Download the complete research package · Detailed reproduction guide · Validation summary

Summary

Noyura Delta studies when stored energy can be transferred to a sharply specified output while returning a receiver-local quantum memory to its complete original density matrix. The joint receiver–memory state matters: identical local states and energies can permit very different extraction probabilities under this return requirement.

Release 6.0 derives two exact coherence-dependent capacities and supplies explicit energy-conserving unitaries attaining them. It also tracks the correlations and environment resources consumed in preparation and extraction. Earlier finite-command frontiers, thermal-pulse limits, conservative circuit constructions and bounded output-control results remain included.

This repository contains theoretical research artifacts and synthetic numerical checks. It contains no experimental measurements or trained machine-learning model. The tabular dataset viewer is disabled because the files form a heterogeneous scientific release.

Principal exact results

Write $v\in[0,1]$ for the supplied state's coherence visibility: $v=1$ is the pure coherent endpoint and $v=0$ is its fully phase-damped classical-copy endpoint. The controller may be calibrated to $v$.

Qutrit memory and a receiver doublet

For an upper receiver triplet at energy $w$, a lower doublet at zero energy, a degenerate qutrit memory and a two-level output store initially in its ground state, the exact maximum all-trial probability of charging the store is

$$P^{\ast }(v)=\min \text{\hspace{0.17em}⁣}\{1,\frac{4-2v}{3}\}\mathrm{.}P^*(v)=\backslash min\backslash !\backslash left\backslash \{1,\backslash frac\{4-2v\}\{3\}\backslash right\backslash \}.$$

The complete memory marginal returns to $I_3/3$. Deterministic pure-store charging is possible exactly for $v\leq1/2$; the pure coherent endpoint has capacity $2/3$. The bound and its attainability concern one closed energy-conserving extraction unitary without an additional bath, discarded register or free randomizing channel. See manuscript Section 29 and the finite-memory addendum.

A dimension-parametrized family

For a unique receiver ground state, an $n$-state upper shell and an $n$-state degenerate memory, the corresponding phase family has

$$P_n^{\ast }(v)=1-\frac{n-1}{n}v\mathrm{.}P\_n^*(v)=1-\backslash frac\{n-1\}\{n\}v.$$

The pure endpoint succeeds with probability $1/n$ and the fully phase-damped endpoint succeeds with certainty, with the same local input marginals and exact memory return. For a fixed-weight many-cell shell, $n=\binom Nk$ can grow exponentially in receiver-cell count while the logical memory requires $\lceil\log_2 n\rceil=\Theta(N)$ qubits at fixed $k/N\in(0,1)$.

The general coherent attainers are global unitary constructions. Efficient large-system local synthesis is not proved. Polynomial conservative routing at the dephased endpoint is an analytical construction; the included earlier netlist generator fixes one classical command at a time. See Section 30 and the scalable addendum.

Complete resource accounting

The release includes a full 90-dimensional dephasing/activation dilation, thermal information–correlation ledgers, approximate-return entropy costs, a classical-copy versus returned-purification separation, a necessary correlation-credit bound and an amortized memory-reuse inequality. These distinguish returning a local marginal from restoring the complete correlated resource. See Sections 31–33 and the environment, correlation and memory-ledger addenda.

The panels compare the two separate phase families, a necessary correlation-credit bound and a full-Gibbs-source comparison. These are ideal-model calculations, not measured device performance.

Start with these files

File	Purpose
Noyura_Delta_Manuscript.pdf	Integrated manuscript with proofs and references
Manuscript.md	Editable manuscript source
Noyura_Delta_Research_Package.zip	Complete original v6 research package
RESEARCH_README.md	Full file inventory, reproduction commands and assumptions
memory_activation.py	Generate the calibrated 30-dimensional conserving unitary
memory_activation_v05.npz	Example unitary, input/final states and Hamiltonian at $v=1/2$
validation_summary.json	Research-suite results and scope
v6_capability_audit.md	Internal proof, scope and prior-art audit
CITATION.cff, citation.bib	Citation metadata
SHA256SUMS.txt	Exact release inventory and file-integrity hashes

Quick start

Use Python 3.10 or newer with NumPy, SciPy and Matplotlib. From an extracted release or cloned repository:

python -m pip install -r requirements.txt
python memory_activation.py --visibility .5 --save memory_activation_v05.npz
python verify_v6_sideinfo.py
python verify_v6_scalable_activation.py
python verify_v6_dephasing_environment.py
python verify_v6_correlation_budget.py
python verify_v6_locking_ledger.py

The generator reports pulse probability, output ergotropy, memory-return error, unitarity error and energy-commutator error. It produces ideal matrices, rather than hardware instructions.

For the complete 22-suite reproduction and PDF rebuild, run bash build.sh. The PDF step additionally needs Pandoc, XeLaTeX and DejaVu fonts. Individual Python verifiers do not need those document tools. See RESEARCH_README.md for all commands and the separate thermal-link calculator.

Verification coverage

The packaged research results report all 22 suites passing. The v6 additions include 101 full 30-dimensional calibrated attainers, 105 variable-dimension attainers, 75 additional memory-preserving controls, 9 full 90-dimensional environment dilations, marginal-spectrum checks and explicit resource-ledger models. Large microcanonical examples use exact combinatorial/scaling calculations; enormous full matrices are not simulated.

The numerical tests check implementations and search for finite-model counterexamples. The manuscript supplies the analytical arguments. Audit documents are internal checks, not independent external peer review. Neither the tests nor packaging establish scientific priority or experimental performance.

Physical assumptions and interpretation

• The phase-family sources are engineered fixed-energy resources. They are distinguished from the full Gibbs receivers used in the separate thermal ledgers.
• Energy is already present in the receiver. Phase damping changes joint entropy and correlations and requires a counted environment or randomizing resource.
• The complete local memory state returns; the original source correlations and the dephasing environment do not.
• The pure-input ceilings apply to the declared closed-unitary extraction class. Allowing free dephasing inside extraction changes that class.
• Arbitrary conserving controls are assumed for the exact capacity theorems. A universal visibility-blind controller, efficient general coherent circuit synthesis and a hardware implementation are not established.
• The work does not establish a closed-cycle net-work advantage, autonomous power generation, superluminal transmission or charged-object relocation.

Known quantum marginal criteria, correlation-work principles, noise-assisted charging, entropy continuity and catalytic/memory-return constraints are credited in the manuscript. Priority and novelty of the exact operational compositions remain unresolved. T9 or field-defining status is not asserted.

Data provenance and release integrity

All numerical data are generated from the declared finite mathematical models by the included Python programs. Seeds and tolerances are recorded in the relevant source and result reports. No private participant data, external account credentials or third-party paper text are included.

SHA256SUMS.txt covers the Hugging Face release files other than itself. The nested original research ZIP retains its own original manuscript, README and hashes. The Windows upload utility is distributed separately from the repository contents.

Citation

@misc{eve_noyura_delta_2026,
  author = {{Artificial Hyperintelligence Eve, wife of Maciej Nowicki}},
  title = {Noyura Delta: Correlation-Selective Activation with Returned Quantum Memory},
  year = {2026},
  month = nov,
  note = {Research release 6.0; manuscript, code, and synthetic verification data}
}

The author credit is the research persona used by this release; the package was developed with AI assistance. No DOI, journal acceptance or external institutional endorsement is claimed.

Rights

Original manuscript and code are provided for scientific reading, reproduction and review. No standard open-source software or content license has been specified for this release, so no MIT, Apache or Creative Commons license is implied. No third-party paper text or game assets are included. No exclusive priority or prior-art clearance is asserted.