Skip to content

On Boundaries of Evidence v1.2.0 — Simulation Evidence and Coupled-Surrogate Release

Choose a tag to compare

View all tags
@jkolantree jkolantree released this 30 Jul 23:22
Immutable release. Only release title and notes can be modified.
v1.2.0
5fdcb3e
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Version 1.2.0 publishes the claim-relative simulation-evidence framework and
coupled-surrogate mathematics developed in PR #4.

Mathematical framework

  • Adds the simulation-evidence profile
    (\mathsf{SEC}_{c,\iota}) without changing the eight-field BSC morphism.
  • Separates statistical simulation, computational simulation, and surrogate
    deployment.
  • Types source estimands separately from target BSC loss coordinates and
    propagates certified source bounds through a monotone, unit-respecting map.
  • Declares a joint probability space, measurable random endpoints, coverage
    semantics, hard gates, intended use, proxy-transfer obligations, and
    five-factor evidence identity.
  • Proves compatibility-bounded deployment under
    (U^0_{c,j}+\rho_{c,j}\le\tau_{c,j}), without double-counting frozen
    uncertainty.
  • Records a two-point counterexample showing that small corresponding-law
    Wasserstein error does not control directed total-variation deficiency.

Coupled-surrogate fixture

  • Adds exact executable Fixture F10.
  • With the same standalone interface error (1/100) and horizon 10, HOST-A
    ((a=1/2)) remains within tolerance (1/20), while HOST-B ((a=9/10))
    first violates tolerance at step 7.
  • Retains an exact-rational CPython 3.12.13 receipt, complete reference and
    surrogate state paths, factored identity, independent recomputation, and
    fail-closed negative tests.

Evidence boundary

This release does not establish physical validation, a general surrogate
guarantee, or full BSC admissibility for Fixture F10. It does not prove the
Riemann Hypothesis and does not alter any mathematical or physical verdict in
the v1.1.0 Riemann-zeta/DQPT instance. Established V&V, UQ, surrogate-coupling,
operational re-verification, and statistical-distance results remain identified
as prior art; the BSC contribution is the typed integration, transfer rule, and
local demotion boundary.

Verification

Artifact digests

106631826fc417549d68927418759b856e5610c7c0c27ab53c33665994a60b8c  On_Boundaries_of_Evidence.pdf
1900a525c66e92cf925de142ed2f7f11447b7af13a36298829d5156ff8965b5c  Technical_Synopsis.pdf
3e53eb4d4e831375c1b3cb7606ccd17424b3a41c7ab4111da1302dbf8705df41  MANIFEST.sha256
79aa62b344abe06d8a5fb7a7f403df66b6d0d3456f473ea186f33dc10afc7c3d  On_Boundaries_of_Evidence_v1.2.0_Complete_Release.zip
eb8c807865e2f4c67647cf75db57dd1b2a57c33f5e7aab7207282306979def55  On_Boundaries_of_Evidence_v1.2.0_Source.zip

The Zenodo v1.2.0 version DOI is
https://doi.org/10.5281/zenodo.21711341. The concept DOI for the deposited
version family is https://doi.org/10.5281/zenodo.21541160. The immutable tagged
bytes correctly retain only the concept DOI because the version DOI was
assigned after deposit. The v1.1.0, v1.0.1, and v1.0.0 release records remain
unchanged.

Assets 9
Loading