A trilogy in three papers, one runnable cognitive instrument, and four frozen experiment tracks.
The public entry point is deliberately small: no model weights, no training pipeline, and no domain labels inside the instrument. The quickstarts below run with Python stdlib only.
git clone https://github.com/JackeyLGene/GBE.git
cd GBE
python docs\quickstart_geruon.py
python docs\quickstart_ee_self.pyExpected first checks:
quickstart_geruon.pyruns a solo Geruon and printstau,F,wit_rate, andarrow norm.quickstart_ee_self.pyruns a 3-cavity Self and prints a cross-harm time series.- No
pip installis required for these two scripts. They are tested with Python 3.8+.
After the quickstarts, read the manuals in order:
| # | File | Purpose |
|---|---|---|
| 1 | docs/GERUON_MANUAL.md | Instrument API, Faraday readings, calibration vocabulary, and minimal usage |
| 2 | docs/EE_MANUAL.md | Self / We / Codex assembly and the four experiment architectures |
| 3 | docs/experiment-passive-calibration-report.md | Passive instrument calibration and frozen readout policy |
Two demo videos generated from the same instrument. Each script also runs locally in under a minute.
demo_self_sine.mp4
A synthetic stream with three known structural boundaries. A 3-cavity Self processes the stream blind — cross-harm peaks at all three boundary positions. The video explains what cross-harm and τ measure, and why they are orthogonal readings.
demo_geruon_bach.mp4
Bach's BWV 846 as 12-dim chroma. A solo Geruon with zero music theory — F responds to harmonic density, wit density tracks structural novelty across four sections, and their divergence shows the two readings are orthogonal. Piano audio synthesized from MIDI.
To regenerate both videos:
python demo\make_video.pyRequires matplotlib, numpy, soundfile, and ffmpeg.
experiments/shepard/_shepard.py · interpretation
The Shepard tone is an auditory illusion of endless ascent. It simultaneously contains a chroma circle (closed) and a height drift (open) — a structural contradiction. A solo Geruon does NOT detect this as a paradox: it locks onto the strong chroma periodicity and barely registers the height drift. A 3-cavity Self confirms the null result. This is an honest negative: the instrument is immune to perceptual illusions that lack physical structure. Good for understanding what the instrument can and cannot see.
GBE is a centroid detector with memory pressure.
The instrument repeatedly applies two operations:
- merge similar frames
- prune frames that are not reinforced
Under finite memory, those operations converge streams into structural centroids. Multiple Geruon cavities coupled through shared fields produce cross-harm: a scalar readout of structural divergence across time lenses. Centroids that survive the cycle can be externalized into a Codex, outlasting the cavity that produced them.
The same core instrument is used across four domains:
| Domain | Experiment role | Core reported result |
|---|---|---|
| UN diplomatic voting | Civilizational structure / forward prediction | 2025 displacement rank 1/79; falsifiable 2026-2028 trajectory |
| Bach WTC | Externalized cognition and Codex convergence | Transposition equivariance 100%; five-stage Codex loop |
| DNA alignments | Evolutionary archive readability | Exon/intron d≈-0.97; AHSG fork-column d=+1.51 |
| RNA transcripts | Molecular operation boundary | Blind CDS stop boundary 98-100% within three windows; start boundary weaker but positive |
| Paper | Role | Main object |
|---|---|---|
| I: GEME | Static self-reference | Shannon-Gödel bridge |
| II: BGM | Time enters the bridge | Temporal decoupling and differentiation |
| III: EE | Externalized cognition | Geruon / Self / We / Codex and four-domain evidence |
- Paper I: GEME — DOI: 10.5281/zenodo.20344974
- Paper II: BGM — DOI: 10.5281/zenodo.20238099
- Paper III: EE — v1.6, May 2026
- EE Supplement
GBE/
├── docs/
│ ├── quickstart_geruon.py
│ ├── quickstart_ee_self.py
│ ├── GERUON_MANUAL.md
│ ├── EE_MANUAL.md
│ ├── experiment-passive-calibration-report.md
│ └── experiment-shepard-paradox.md
├── demo/
│ ├── demo_self_sine.mp4 # sine boundary detection (3-cavity Self)
│ ├── demo_geruon_bach.mp4 # Bach prelude (solo Geruon, with audio)
│ ├── demo_self_sine.py # text-mode demo (runnable)
│ ├── demo_geruon_bach.py # text-mode demo (runnable)
│ ├── make_video.py # regenerates both MP4s
│ └── bwv846.mid # Bach C major prelude
├── code/
│ ├── geme.py
│ ├── geruon.py
│ ├── we_core.py
│ ├── bgm_core.py
│ ├── bgm_bacteria.py
│ ├── midi_encoder.py
│ └── _calibrate_*.py
├── experiments/
│ ├── un/
│ ├── wtc/
│ ├── dna/
│ ├── rna/
│ └── shepard/
└── paper/
├── gEME.pdf
├── bGM.pdf
├── ee.pdf
└── ee_supplement.pdf
- The two quickstarts are zero-dependency checks for the public instrument.
- The
demo/andexperiments/shepard/scripts are small runnable bridge examples. - The frozen experiment scripts are included under
experiments/. - Large raw datasets are not committed; data sources and expected inputs are documented in each experiment track.
- Domain experiments may require packages such as
numpy,pandas,pyarrow, orpybigtools, depending on the script.
The fastest review path is: run both quickstarts, skim the Geruon manual, then inspect one experiment track.