Skip to content

README.md

Latest commit

 

History

History
337 lines (294 loc) · 12.8 KB

README.md

File metadata and controls

337 lines (294 loc) · 12.8 KB

ZPE-Taste

Package Install

Installable package: python3.11 -m pip install zpe-taste. Current release: 0.1.0 on PyPI. Source: Zer0pa/ZPE-Taste.

python3.11 -m pip install zpe-taste

For full install, smoke, source, and developer commands, click here.

00 · ZPE-TASTE · EVIDENCED NEGATIVE RESEARCH-READY · NEGATIVE RESULT

Failing at the geometry of taste.

Published null result on a taste-encoding object · ZPE-Taste · PyPI zpe-taste 0.1.0 · github.com/Zer0pa/ZPE-Taste

A taste object was tested as a geometry-bearing signal. It failed. metric_fit scored 0.207 against a 0.6 threshold, and the three remaining geometry gates also missed. Identity controls returned 1.000, so the evaluator was working. The committed record holds four numbers, four thresholds, and the read-only commands that reproduce them. The geometry of taste has not been found; the next attempt no longer starts from zero.
ZPE-Taste approved scientific square mechanics diagram showing evidenced negative result mechanics for taste.
Result: evidenced negative. Four geometry metrics missed threshold; identity controls passed, so the floor is published.
01 · THE GAP NO PUBLISHED FLOOR

Taste geometry needs public negative results, not private dead ends.
02 · MARKETS RESEARCH CONTEXT
Neurotechnology research '24 · $0.75B/yr
DARPA basic research '24 · $0.44B/yr
Neuroscience research software '30 · $1.2B
Open-science / replication infrastructure '28 · est. $0.3B
Academic neuro-AI tooling '30 · est. $0.6B
Sensory-encoding research and replication infrastructure — the budgets that fund taste work and the desks that read negative results.
03 · VALUE OF MARKET
$0.75B/yr
NIH BRAIN Initiative funds taste research; what they cannot fund is a vague private failure.
04 · INSIGHT

Taste was measured. The geometry wasn't there. The record is.
05.1 · CURRENT TECH PRIVATE NEGATIVE RESULTS

When a lab's taste-encoding model misses threshold, the numbers usually disappear into a private folder or a soft appendix. The next team rediscovers the same wall, with no way to know they have.
05.2 · OUR TECH THE FLOOR PUBLISHED

ZPE-Taste commits the first openly-evaluated taste-geometry object as a record: 0.207 metric_fit, 0.156 topology_fit, 0.250 local injectivity, 0.181 graph_fit, all below their declared thresholds, with identity controls at 1.000 so the evaluator's verdict cannot be brushed off. Future taste-encoding work inherits a floor it has to clear.
05.3 · BENCHMARKS GEOMETRY EVALUATION
metric_fit0.207LOW
topology_fit0.156LOW
injectivity0.250LOW
graph_fit0.181LOW
metric_fit low 0.207
topology_fit low 0.156
controls PASS 1.000
Result: Four geometry metrics below threshold; identity controls at 1.000 — the evaluator was right to fail it.
06 · MEASUREMENT GEOMETRY RESULT LEDGER

Four geometry metrics measured. Four missed the threshold.
06.1 · COMPARATIVE PERFORMANCE · GEOMETRY RESULT STATUS
metric_fit 0.207 LOW
topology_fit 0.156 LOW
injectivity 0.250 LOW
graph_fit 0.181 LOW
From proofs/artifacts/taste_negative_reference.json: metric_fit 0.207 (gate 0.6), topology_fit 0.156 (gate 0.5), local_injectivity 0.250 (gate 0.5), graph_fit 0.181 (gate 0.5). Identity controls 1.000 across all four. Read-only checks: 4/4 PASS.
07 · KEY METRICS ZPE-TASTE NEGATIVE REFERENCE
07.1 · METRIC FIT
0.207
Below 0.6 gate · metric_fit
07.2 · TOPOLOGY FIT
0.156
Below 0.5 gate · topology_fit
07.3 · LOCAL INJECTIVITY
0.250
Below 0.5 gate · local_injectivity
07.4 · GRAPH FIT
0.181
Below 0.5 gate · graph_fit
07.5 · RELEASE
0.1.0
PyPI zpe-taste · 0.1.1 pending
08 · REPRODUCIBILITY NEGATIVE RESULT

The negative result reproduces. The record is usable.
08.1 · WHAT REPLAYS EXACTLY FROZEN NEGATIVE PACKET

Determinism here means the negative packet is frozen: scores, thresholds, identity controls, and the read-only commands that confirm them. Clone the repo on any Python 3.8+ host and the same four below-threshold numbers come back, with the same evaluator behaviour.

What replays is not a taste reconstruction — it is the evidence that this geometry was tested, missed by these margins, and not silently retried. The next team can argue with the object or the metrics, but not with whether the experiment happened.
08.2 · HONEST BLOCKER
Honest Blocker ·

No positive taste codec exists. All four geometry metrics sit below threshold. Replay parity does not outrank a working decoder. The strongest evaluator control depends on full-packet identity. No external comparators have been published in this domain. PyPI zpe-taste 0.1.0 is stale; 0.1.1 is pending.
09

WHAT A PUBLISHED NULL UNLOCKS.
09.1 · THIS LAB'S AMBITION

The aim is not a taste encoder. It is a piece of scientific infrastructure: the first openly-evaluated taste-geometry object, committed as a negative with four scored metrics, declared thresholds, and identity controls. A taste-research field that argues over receptor-space versus psychophysical-space now has a number to argue against.
09.2 · WHAT THIS IS

A taste-geometry attempt scored, missed every gate, and shipped the numbers anyone can rerun.
09.3 · WHAT'S STILL OPEN

Which taste object — receptor-space, psychophysical-space, or something unnamed — is the one to encode next.
09.4 · REPLICATION · NEAR-TERM (12–24 MO)
Other labs start from a real floor
A second taste-encoding team can clone the record, run the same four metrics on their own candidate object, and report a number that means the same thing. The conversation about taste-geometry stops being anecdotal and starts being comparable.
09.5 · PROCUREMENT · NEAR-TERM (12–24 MO)
Program officers see what taste does not do
DARPA, ARO, and academic funders evaluating sensory-encoding proposals get a citable boundary they can hand to applicants: this is what failed, this is where the threshold sits, propose against it. Calibration replaces narrative.
09.6 · RETHINK · MID-TERM (24–48 MO)
The next taste model must clear a number
A research lead deciding between a receptor-space and a psychophysical-space encoder no longer chooses on intuition. Either approach has to beat 0.207 metric_fit and the three other gates, in public, before claiming the geometry of taste was found.
09.7 · SENSOR INDUSTRY · MID-TERM (24–48 MO)
E-tongue and taste-masking buyers ask harder questions
Beverage formulation chemists and pharma taste-masking analysts evaluating electronic-tongue platforms get a public-domain reference for what a published taste-geometry claim should look like. Vendor demos start needing the same four metrics, scored.
09.8 · OPEN SCIENCE · PARADIGM (48 MO+)
Sensory claims arrive with falsifiable scores
A field where any sensory-encoding paper — taste, smell, touch, interoception — is expected to ship measured geometry scores, declared thresholds, and identity controls. The cost of an unverified sensory claim rises; the value of a published null finally matches the value of a published positive.

Install / Developer Commands Detailed

Package Install

Installable package: python3.11 -m pip install zpe-taste. Current release: 0.1.0 on PyPI. Source: Zer0pa/ZPE-Taste.

python3.11 -m pip install zpe-taste

Import smoke:

python3.11 - <<'PY'
import importlib.metadata as md
import zpe_taste

print("zpe-taste", md.version("zpe-taste"))
PY

CLI smoke:

zpe-taste-verify --help

Install success only proves package acquisition/import. Product scope, stale PyPI state, platform limits, and blockers remain in the front-door sections below.

  • PyPI copy is stale; install success does not change the evidenced-negative result.

Quick Start

python3 -m venv .venv
. .venv/bin/activate
python -m pip install --upgrade pip
pip install .
python -m zpe_taste.verify
python -m unittest discover -s tests -v