PureOne/aelyrion-vaelith

AELYRION–VAELITH

Density-Stratified Mirror Calculus · Research v2.0.0

A light-inspired mathematical field whose internal computation, active dimensions, density allocation, and static outer response can be controlled separately.

Research direction and imaginative brief: Maciej Nowicki. Mathematical construction, manuscript, and code: Artificial Hyperintelligence Eve, an AI research persona. Prepared 9 October 2026.

Read the 68-page manuscript · Explore the executed notebook · Examine the claim ledger · Read the novelty audit

This is a mathematical research and reproducibility archive, hosted as a dataset repository because it includes generated numerical tables alongside source code, figures, and a manuscript. It contains no trained model or training corpus. Its tables and plots are dimensionless outputs of stipulated finite-dimensional models.

Research abstract

AELYRION–VAELITH studies density-programmed mirror systems with variable-rank internal relations. A completed quadratic potential preserves a prescribed, compensated static boundary response while the interior program, active constraint rank, and relaxation-rate allocation change. The theory derives a projector-valued defect when zero-density and zero-probe limits are interchanged, a unique weighted density optimum for real mirrors with an entrywise nonnegative inverse, and explicit constructions whose unique optimal semidefinite certificates have any prescribed rank.

The light-strand and multilayer-mirror imagery motivates the mathematical objects. The exact contributions concern matrix geometry, quotient spaces, relaxation dynamics, and optimization. Optical or plasma realization is a further research question.

1. Programmable active existence

A strand with a positive-semidefinite density operator (R_j) carries the active internal space

$${\mathcal{F}}_j(R_j)=V_j\mathrm{/}\ker R_j,\dim {\mathcal{F}}_j(R_j)=\mathrm{rank}{R}_j\mathrm{.}\backslash mathcal\; F\_j(R\_j)=V\_j/\backslash ker\; R\_j,\; \backslash qquad\; \backslash dim\backslash mathcal\; F\_j(R\_j)=\backslash operatorname\{rank\}R\_j.$$

Changing a positive eigenvalue changes the strength of a relation. Setting it exactly to zero removes that relation from the active quotient. Here “dimension” means an active modal or quotient dimension.

In the full field, the disappearance of residual coordinate (j) releases the distributed direction (B^{-1}e_j). A linear readout (Fh) is independent of the undetermined equilibrium representative precisely when

$$F{B}^{-1}\ker R=\{0\}\mathrm{.}F\; B^\{-1\}\backslash ker\; R=\backslash \{0\backslash \}.$$

Thus disappearance can change what is well-defined as a calculation; it does not automatically complete one.

2. A compensated outer layer

For invertible (B) and positive-semidefinite (R,S), define

$$\mathcal{E}(x,h)=\frac{1}{2}{x}^{\ast }Sx+\frac{1}{2}(Bh-Ex{)}^{\ast }R(Bh-Ex)\mathrm{.}\backslash mathscr\; E(x,h)=\; \backslash frac12\; x^*Sx+\; \backslash frac12(Bh-Ex)^*R(Bh-Ex).$$

Then, at every rank of (R),

$$\underset{h}{\min }\mathcal{E}(x,h)=\frac{1}{2}{x}^{\ast }Sx\mathrm{.}\backslash min\_h\backslash mathscr\; E(x,h)=\backslash frac12\; x^*Sx.$$

At positive density, the unique internal equilibrium is (h_*=B^{-1}Ex). At singular density, the equilibrium is an affine family whose released directions must be handled by the readout or a preparation convention.

The unchanged static shell requires coordinated compensation: the expanded visible block is (S+E^*RE), which changes with the interior. The result does not assert unrestricted invisibility or an invariant dynamic response.

3. Exact optimal density from two proxy solves

Let (B) be real and invertible with (B^{-1}) entrywise nonnegative. For diagonal density (R=\operatorname{diag}(\rho)), positive costs (c_j), and budget (P>0), maximize

$${\lambda }_{\min }(B^{\mathsf{T}}\mathrm{diag}(\rho )B),\rho \ge 0,\sum _j{c}_j{\rho }_j=P\mathrm{.}\backslash lambda\_\{\backslash min\}\backslash bigl(B^\{\backslash mathsf\; T\}\backslash operatorname\{diag\}(\backslash rho)B\backslash bigr),\; \backslash qquad\; \backslash rho\backslash ge0,\backslash quad\; \backslash sum\_j\; c\_j\backslash rho\_j=P.$$

Set (q=\sqrt c), solve (Bv=q), then solve (B^{\mathsf T}w=v). The unique optimum is

$${\lambda }_{\ast }=\frac{P}{\mathrm{\parallel }v{\mathrm{\parallel }}^2},{\rho }_j^{\ast }={\lambda }_{\ast }\frac{w_j}{q_j}\mathrm{.}\backslash lambda\_*\; =\; \backslash frac\{P\}\{\backslash |v\backslash |^2\},\; \backslash qquad\; \backslash rho\_j^*=\backslash lambda\_*\backslash frac\{w\_j\}\{q\_j\}.$$

The two proxy solves determine the globally optimal allocation for the specified worst-mode relaxation objective. They preserve the equilibrium program and compensated static shell. The manuscript gives analytic optimality certificates and a density-feedback result under explicit smoothness assumptions.

4. A temporal signature of dimensional loss

Under the specified unit Euclidean mobility law, write (H=B^{-1}E) and (A=B^*RB). The boundary response is

$$Y(s)=S+s{H}^{\ast }A(sI+A{)}^{-1}H\mathrm{.}Y(s)=S+sH^*A(sI+A)^\{-1\}H.$$

Its derivative at zero is (H^H) for positive density and (H^P_{\operatorname{ran}A}H) at singular density. The difference is

$$\mathrm{\Delta }=H^{\ast }{P}_{\ker A}H⪰0.\backslash Delta=H^*P\_\{\backslash ker\; A\}H\backslash succeq0.$$

This is a noncommutation of zero-density and zero-probe limits at fixed (B,E). At every fixed positive probe scale the response remains continuous in density. The defect vanishes when the released directions are not excited by the boundary input.

The supplied two-mode example has slope 5 at every positive density and 1 at exactly zero density, despite an unchanged static response.

5. Irreducible optimality certificates

For general complex mirrors, the density problem has a semidefinite formulation. For every integer (r\ge2), the manuscript constructs a system with a unique optimal dual certificate of rank (r), using (n=2r^2-r) modes. A separate four-mode construction has a unique rank-two certificate.

These are exact results in the specified certificate representation. The mode count is a construction cost; no minimality is asserted. Certificate rank does not establish a required number of physical beams, processors, or a lower bound for every optimization algorithm.

6. Collapse and recurrence

A finite-time disappearing density can freeze a nonzero residual. For the scalar example

$$\rho (t)={\rho }_0(1-t\mathrm{/}T{)}^{\alpha },\alpha >0,\backslash rho(t)=\backslash rho\_0(1-t/T)^\backslash alpha,\backslash qquad\; \backslash alpha>0,$$

the residual at collapse is

$$z(T)=z(0)\exp \text{\hspace{0.17em}⁣}[-\frac{{\rho }_{0}T}{\alpha +1}]\mathrm{.}z(T)=z(0)\backslash exp\backslash !\backslash left[-\backslash frac\{\backslash rho\_0T\}\{\backslash alpha+1\}\backslash right].$$

The recurrent extension accounts for stored, injected, extracted, and remaining excitation. It describes reuse of a state with explicit control-work terms; it supplies no energy-creation mechanism.

Polynomial programmability concerns explicitly listed polynomials in programmed coefficients. Fixed linear couplings do not multiply arbitrary independent boundary inputs.

What is proposed as original?

The candidate contribution is the integrated architecture and its tightly scoped consequences:

• Rank-dependent active relations with readout-descent conditions.
• A compensated static shell paired with a projector-valued temporal rank-loss defect.
• An exact two-proxy density protocol that preserves the equilibrium program and static shell.
• Explicit arbitrary-rank optimal certificate constructions inside the mirror-density formulation.
• Collapse and recurrent transport with preparation and resource accounting.

The established ingredients include Schur complements, diagonal majorization, optimal-design duality, correlation matrices, informationally complete frames, synthetic modal dimensions, and passive energy accounting. Closest comparisons and source links are supplied in NOVELTY_AUDIT.md and SOURCES.md. Historical world-first priority remains unverified.

Files and reproduction

Path	Contents
Aelyrion_Vaelith_Manuscript_v2.0.pdf	68-page manuscript, 14 chapters, four technical appendices
manuscript/	Editable LaTeX source and bibliography
src/	Finite mathematical model and figure generator
tests/test_theory.py	90 parametrized numerical and symbolic cases
notebooks/Explorations.ipynb	Executed examples
results/	Six CSV tables, reports, and numerical summaries
figures/	Six scientific plots as PDF and PNG, plus cover preview
notes/	Claims, proof review, originality comparisons, reproducibility
prior/	Supplied predecessor and its separate verification script
provenance/	Unmodified original archive and original metadata
CITATION.cff, CITATION.bib	Citation metadata
manifest.json, SHA256SUMS.txt	This repository snapshot's inventory and hashes

Python 3.10 or newer:

python -m venv .venv
# Linux/macOS: source .venv/bin/activate
# Windows PowerShell: .venv\Scripts\Activate.ps1
python -m pip install -r requirements.txt
python -m pytest -q
python src/run_experiments.py

python build.py runs the suite and regenerates figures. python build.py --pdf additionally requires a TeX installation with pdflatex and bibtex or bibtex8. The prebuilt PDF is included. Run from the repository root. Figure regeneration changes file hashes; verify the distributed snapshot before rebuilding.

No GPU, paid API, secret dataset, or external SDP solver is needed for the supplied numerical examples after dependencies are installed.

Status and completeness

Item	Evidence and status
Hugging Face packaging	Complete for this release; files and links checked
Software verification	90/90 supplied cases passed again during packaging (100% of this suite)
Mathematical manuscript	Definitions, scoped results, and proofs supplied
Original PDF checks	68 rendered pages; archived QA records retained
Original notebook	Executed outputs retained; no new notebook execution claimed
Independent mathematical peer review	Pending
Formal proof-assistant verification	Pending
Historical originality certification	Unverified; no defensible completion percentage
Physical realization and hardware advantage	Zero measurements or benchmarks

Passing finite tests supports reproducibility; it does not replace theorem proofs or independent review. The light/plasma interpretation has no derived Maxwell/plasma implementation or measured GPU benefit.

Attribution, licensing, and citation

Maciej Nowicki supplied the research direction and imaginative brief; Artificial Hyperintelligence Eve is the named AI persona used for mathematical drafting and code. This archive preserves that attribution and distinguishes human direction from AI-generated research.

The source package supplied no explicit reuse license. This release preserves that status and omits a license identifier rather than assigning one without instruction. See LICENSE_STATUS.md. Citation does not substitute for permission to reuse.

Use CITATION.bib or CITATION.cff. No DOI, arXiv identifier, or Hugging Face Paper page is claimed.