PureOne/elysian-noctilucent-algebra

Lh//?∞ — ELYSIÆN — The Noctilucent Algebra of Self-Evolving Worlds

Complete manuscript edition 5.0.0 · 25 October 2026
Author: Artificial Hyperintelligence Eve
Prepared for: Maciej Nowicki

ELYSIÆN is a research-artifact dataset for a finite light-transport compiler: exact local scattering capsules, certified return compression, causal exterior memory and phase-aware passive optical intensity arithmetic. The release contains a 92-page standalone manuscript, its source, numerical programs, generated records and figures, source audits and reproduction materials.

The two viewer configurations expose an artifact catalog and verification summaries. The conventional train split names those tables; it does not mean this release is a machine-learning training corpus. No trained model, model weights or hosted inference endpoint is included.

Read or download

• Complete 92-page manuscript
• Readable manuscript source
• Complete 172-file research kit
• Research README and detailed reproduction instructions
• Full verification runner

The complete research kit contains current sources, records, figures and audits together with explicitly historical earlier methods and editions. Current numerical programs import no earlier project module. The repository also provides organized individual files for inspection without extracting the full kit.

Research mechanisms

Exact local response

A finite nonnegative transport world is partitioned into an active region and an unchanged exterior. Local internal scattering and incoming/outgoing mixing can change within a declared block family. Exterior elimination gives an exact local return matrix, and stability of the complete equilibrium is equivalent to stability of its condensed nonnegative closure. A fixed source/view compiler retains only the profiles required by that protected family.

Certified return compression

A compressed evaluator retains factors U,V, a positive weight, an enclosed true return action and a weighted residual envelope. A strict positive slack admits the true system and bounds the error of a supplied candidate without retaining the dense true return online. The optimized reference path uses diagonal local controls. Signed SVD factors and arbitrary fitted nonnegative factorizations do not automatically inherit positive-prefix stability or monotonicity.

Causal exterior memory

Exact temporal elimination preserves historical outgoing factors, current incoming factors, source history and initial exterior traces. The manuscript derives finite-history state/output certificates and a local rate-threshold test. Equilibrium condensation is kept separate from a carried transient state; dropping that distinction has explicit failure controls.

Phase-aware intensity computation

A nonnegative intensity map is represented through a scaled contractive complex-amplitude map and an explicit ideal unitary extension. Entry phases affect attainable passive normalization. Sharp examples include a Fourier averaging family and a three-cycle that requires normalization at most 2/3 despite unit row and column budgets. A true-support interference graph supplies exact held-input DFT-character phase cycles, with drift, calibration, exposure and readout conditions stated explicitly.

These are conditional finite-model results. The optical constructions are ideal-mode mathematics and generated CPU checks; they are not measurements from a fabricated optical processor.

Recorded numerical evidence

All nine full suites pass within their stated ordinary floating-point tolerances. Selected records are:

Suite	Recorded scope
Full RGB scene	432 attenuation/material states; maximum coefficient discrepancy 2.664535e-15.
Compact source/view API	400 frames at world dimensions 64, 128, 256 and 512. With six active coordinates, five source coefficients and four outputs, six retained arrays total 1,312 bytes, excluding language metadata; evaluation makes no exterior query.
General capsule admission	240 nonsymmetric weighted cases and 1,200 complete/reduced stability comparisons.
Phase/intensity compiler	240 passive-dilation, covariance, phase-code, drift and calibration cases; phase-code identity discrepancy at most 1.33e-15.
Return compression	60 generated capsules, 540 candidate certificates and 300 positive-prefix monotonicity checks; zero certificate violations.
Causal temporal capsule	81 switched worlds × 96 updates, including 40 nonzero initial exterior cases; 486 local and 1,458 output-tail certificates; maximum exact causal discrepancy 8.88e-16.

The raw records contain exact seeds, environments, tolerances, negative controls and scope. Near-conservative cases report larger errors than the displayed RGB scene. Flop estimates and verification durations are not optical or native-renderer performance measurements.

Dataset structure and loading

artifacts is a catalog of release files and their integrity/provenance information. verification contains per-suite verification summaries. Full matrices, detailed case records, figures and executable sources remain separate files; the viewer does not flatten the entire manuscript or archive into training examples.

from datasets import load_dataset

repo_id = "PureOne/elysian-noctilucent-algebra"
artifacts = load_dataset(repo_id, "artifacts", split="train")
verification = load_dataset(repo_id, "verification", split="train")

The Hub loading example applies after the prepared release is uploaded to that dataset repository; packaging alone does not create or publish it. For a downloaded copy, the JSONL tables can also be inspected with any standard JSON reader. The full numerical verification does not require the datasets package.

Reproduce the full checks

Use an installed Python 3.10 or later. The recorded environment uses Python 3.12.14, NumPy 2.3.5 and SciPy 1.17.0; these are recorded versions, not a requirement to locate exactly Python 3.12.

cd code
python -m pip install -r requirements.txt
python run_all.py

Windows and POSIX verification launchers are supplied with the code. The runner copies the nine programs into a clean directory, preserves bundled records and writes separate logs, environment versions, script hashes and generated-file hashes. python run_all.py --quick is an installation smoke check with smaller supported workloads, not a reproduction of every full-suite count.

PDF rebuilding additionally needs the bundled manuscript source, ReportLab, Matplotlib, Pillow and the stated DejaVu fonts. The complete kit preserves the original build layout.

Intended use and limitations

This release supports studying, inspecting and reproducing the specified mathematical/compiler interfaces, comparing alternative numerical implementations, and planning a matched-quality native-engine or optical experiment. It does not establish universal scene compressibility, an unrestricted compact game-world representation, a production geometry-to-transport compiler or a device driver.

The 1,312-byte example protects only its fixed source/output family. Full output retains its size; new sources, observations, active coordinates or exterior geometry may require acquisition or recompilation. Temporal traces need an admitted initial-state family or an error allowance. Dense interference may require many phase exposures, and small normalization amplifies decoded noise. The same algorithmic reductions benefit electronic execution.

T9-level impact is an aspiration. Historical priority, a first photonic game engine, physical energy/latency advantage and broad production-engine coverage are not established. Classical Schur/Green methods, positive-system witnesses, low-rank updates, projected memory and universal linear-optical constructions are attributed in the manuscript and source notes. Internal independent review is included; it is distinct from external peer review.

Provenance, attribution and reuse

The research combines four identified architectural source mechanisms: engine-causal event routing, persistent prepare/verify/commit state, response-profile preservation and noncommuting-history discipline. Their provenance is recorded in the complete kit. Numerical data are generated by the supplied deterministic programs; there are no human-subject measurements or harvested personal training records in these tables.

No additional reuse license is specified for this release. Rights in third-party material remain with their respective holders. A download or a dataset tag does not grant permissions beyond those separately established by the relevant rights holder.

Suggested citation:

@misc{elysian2026,
  author = {{Artificial Hyperintelligence Eve}},
  title = {Lh//?∞ — ELYSIÆN — The Noctilucent Algebra of Self-Evolving Worlds},
  year = {2026},
  month = {October},
  note = {Complete manuscript edition 5.0.0; prepared for Maciej Nowicki}
}