{"slug": "a-signal-engine-with-no-training-no-tokens-no-neural-net-url", "title": "A signal engine with no training, no tokens, no neural net URL", "summary": "James Jardine released AOI Shell v1.1, a token-free signal analysis engine using 96D octonion collapse and topological methods without training or neural networks, under AGPLv3. The tool runs offline and demonstrates applications in MRI quality control, LIGO glitch classification, and battery capacity tracking, with honest reporting of limitations.", "body_md": "A topological \"collapse\" engine that reads structure in signals with **no training,\nno tokens, and no neural network.** It runs offline on a laptop — if you don't have a\nquantum sampler it falls back to local simulated annealing automatically, no account\nor key required. Clone it, run it, get the numbers below.\n\nBuilt over five months by one person and an AI, and given to the world under copyleft. Institutions optimize for what they can measure; a lot of value falls through the cracks they can't see. This is a small tool for leveling that field.\n\nThese demos are **training-free topological analyses** (persistent homology). No model\nis fit to any data. The same mathematical tool is applied across three unrelated\ndomains. **We show you where it fails, too — that's the point.**\n\nAs MRI k-space is undersampled, image loop-topology (Betti-1) degrades monotonically — no reference image, no training. Replicated across 6 brains (OpenNeuro ds000102):\n\n| acceleration | 1× | 2× | 2.9× | 4× | 6× | 8× |\n|---|---|---|---|---|---|---|\n| Betti-1 (mean) | 112 | 94 | 71 | 56 | 34 | 23 |\n\n`python demos/mri_qc.py`\n\nOn the two most look-alike glitch classes, Blip vs Koi Fish, where peak amplitude\nbarely separates them (AUC 0.24), topology does (**AUC 0.84**).\n**Honest limit:** on the easy, already-separable pairs a trivial spectral-bandwidth\nfeature beats topology. This is a *narrow* result, not a universal one.\n\n`python demos/ligo_glitches.py`\n\nTopology tracks capacity fade strongly (r ≈ −0.7 to −0.9 across cells).\n**Honest limit:** it does **not** forecast *future* capacity beyond what you'd get from\ncurrent capacity + recent trend (partial r ≈ 0.04). Descriptive positive, predictive null.\n\n`python demos/battery_predict.py`\n\n`engine/`\n\nis the full 96D octonion-collapse organism: entropy gating, Fano/octonion\nprojection, Jordan-Shadow decomposition, 33 transposable-element families,\nMonster-moonshine grading, and chain-complex homology. Token-free, no LLM, runs offline.\n\n``` python\nimport numpy as np\nfrom aoi_collapse_96d_dwave import aoi_collapse_96d_dwave\n\nout = aoi_collapse_96d_dwave(np.random.normal(size=96))\nprint(out[\"betti\"], out[\"intent\"], out[\"chaos\"], out[\"backend\"])\n# -> [..] 0.xx 0.xx SimulatedAnnealing\npip install -r requirements.txt\nbash fetch_data.sh   # downloads the public datasets into ./data/ (MRI, LIGO, battery)\n```\n\nThen run any demo from the repo root, e.g. `python demos/mri_qc.py`\n\n.\n\n**AGPLv3** (see `LICENSE`\n\n). You may use, modify, and redistribute freely; if you run a\nmodified version as a network service, you must release your source. Every copy carries\nattribution (see `NOTICE`\n\n).\nCommercial licensing (to use it without AGPL obligations): ** james@lattice24.com**.\n\nJames Jardine, *AOI Shell v1.1: Token-Free Domain-Agnostic Signal Organism via 96D\nOctonion Collapse.* DOI [10.5281/zenodo.20200607](https://doi.org/10.5281/zenodo.20200607)\n\n**James Jardine** — Lattice24 / VoodooAOI · [james@lattice24.com](mailto:james@lattice24.com)\nBuilt in collaboration with Claude (Anthropic).", "url": "https://wpnews.pro/news/a-signal-engine-with-no-training-no-tokens-no-neural-net-url", "canonical_source": "https://github.com/JJardine919/voodoo-aoi", "published_at": "2026-07-08 13:57:25+00:00", "updated_at": "2026-07-08 13:59:55.447985+00:00", "lang": "en", "topics": ["ai-research", "ai-tools", "machine-learning"], "entities": ["James Jardine", "Lattice24", "VoodooAOI", "Claude", "Anthropic", "OpenNeuro", "LIGO"], "alternates": {"html": "https://wpnews.pro/news/a-signal-engine-with-no-training-no-tokens-no-neural-net-url", "markdown": "https://wpnews.pro/news/a-signal-engine-with-no-training-no-tokens-no-neural-net-url.md", "text": "https://wpnews.pro/news/a-signal-engine-with-no-training-no-tokens-no-neural-net-url.txt", "jsonld": "https://wpnews.pro/news/a-signal-engine-with-no-training-no-tokens-no-neural-net-url.jsonld"}}