CHANGELOG.md

Hypergraph Lumpability and Bayesian Refinement

CHANGELOG.md

Download original file

Original Markdown with line links

0.1.4 | September 2026

Changelog

0.1.3 | September 2026

  1. Completed a final proofreading pass across the manuscript, captions, generated tables, release notes, and audit records.
  2. Rewrote dense or mechanical passages in more natural academic prose without changing the main theorem chain.
  3. Removed semicolon punctuation, dash punctuation, and word hyphens from ordinary manuscript prose. Mathematical minus signs, internal labels, program options, proper names, and bibliographic ranges remain where required.
  4. Revised the abstract, section transitions, proof explanations, worked examples, claim boundaries, and conclusion for clearer rhythm and more direct wording.
  5. Clarified that the coefficient comparison in polynomial descent concerns a polynomial with values in the coarse space.
  6. Refined the proof of best coarse interaction attainment so that existence and possible nonuniqueness are stated separately.
  7. Added the compatible initial condition and the sequence evidence definitions to the exact evidence theorem.
  8. Corrected the hidden Markov composition notation and the nonlinear quotient diagram.
  9. Distinguished exact arithmetic from tolerance based approximation in the coordinate algorithm.
  10. Defined the combined quotient and pruning defect before using its norm bounds.
  11. Qualified the likelihood oscillation criterion and stated the probability vector assumption in the evidence error proof.
  12. Expanded the manuscript audit to check these corrections directly.
  13. Added deterministic manuscript checks for proof coverage, version consistency, reference labels, prose punctuation, and known prose regressions.
  14. Executed 33 deterministic checks with 0 failures.
  15. Executed 8 independent pytest functions with 0 failures.
  16. Passed 22 manuscript audit checks with 0 failures.
  17. Rebuilt and visually inspected the 50 page PDF with no undefined references, undefined citations, or overfull boxes.

0.1.2 | September 2026

  1. Strengthened analytic interaction descent with an explicit normal summability hypothesis on every smaller ball.
  2. Sharpened the normalization perturbation inequality by using the larger of the two positive masses in the denominator.
  3. Updated the implementation and tests to use the same normalization bound.
  4. Restricted the likelihood perturbation example to the strictly positive regime.
  5. Added explicit proofs to every corollary that had previously relied on an immediate but unstated derivation.
  6. Added deterministic checks for the symmetric normalization inequality and the positive likelihood regime.
  7. Rebuilt the manuscript as a 50 page internal completion release.

0.1.1 | September 2026

  1. Reconstructed the complete downloadable release after the earlier archive failed to persist.
  2. Restored the 49 page PDF and editable LaTeX manuscript.
  3. Added a clean priq reference package and independent pytest suite.
  4. Regenerated five vector PDF figures, five PNG figures, and machine readable certificate ledgers.
  5. Executed 31 deterministic checks with 0 failures.
  6. Executed 7 independent pytest functions with 0 failures.
  7. Corrected one malformed LaTeX control character and two missing notation macros.
  8. Removed all overfull boxes from the manuscript build.
  9. Added build, status, source crosswalk, manifest, and checksum documentation.

Other PDFs for this work

Original file and source record

This copy matches the recorded source file. Its publication here does not change the work’s review status.

SHA-256: bd8047da67ba21613c05d2e53dc2ddd97351a1313f7f50b9eb68f35cb4d91fc4

Original source on GitHub

Back to Hypergraph Lumpability and Bayesian Refinement