Expanded research archive
Version 1.1.0 extends the public research archive while preserving the
prospectively frozen weighted sparse-Ising confirmation and its primary
result tables unchanged.
Added
- The exhaustive temperature-dependent tree-geometry supporting study,
including its frozen protocol, source instances, compact canonical
summaries, scientific report, reproducible source, and validation records. - The negative operational result that retained target-state cooling power is
not validated as a compressed-state tree-selection objective on the tested
development corpus. - A complete claim-to-evidence map for the new supporting result.
- Release-specific citation, reproducibility, and machine-readable metadata.
- A changelog and direct release navigation from the repository landing page.
- Additional regression checks for citation paths, release metadata, and
packaged validation-record freshness.
Primary confirmed result
The original prospectively frozen result remains unchanged:
- 24 separately generated weighted sparse-Ising targets;
- 1,440 confirmatory optimization trajectories;
- matched chain, random target tree, maximum-weight target tree, and full-graph
representations; - exact preparation-resource accounting;
- complete raw evidence, statistical analysis, and validation records.
At matched width-one resources, the maximum-weight target-supported tree
improves finite-budget optimization over both the chain and the prespecified
random target-supported tree. The full graph remains the trainability ceiling
at a larger exact preparation-description cost.
Interpretation boundary
The new temperature-dependent tree study is supporting developmental evidence,
not an additional confirmation. It records that a locally appealing
cooling-power objective does not improve compressed Gibbs-state quality on its
tested corpus.
No result in this release claims quantum advantage, favorable asymptotic
scaling, universal optimality of the maximum-weight tree, hardware-efficient
preparation, or a validated adaptive tree-selection method.
Citation
For the complete archive, cite release v1.1.0.
For the primary confirmed MAXJ result in its original release state, release
v1.0.0 remains available.
Copy-ready APA-style and BibTeX formats are provided in CITATION.md, with
machine-readable metadata in CITATION.cff.
Validation
The release commit passed:
- Python 3.10, 3.11, 3.12, and 3.13;
- unit, layout, Markdown-math, citation, and link checks;
- package construction;
- primary scientific validation;
- all supporting-study validators.
Large generated intermediates from the exhaustive temperature-tree calculation
remain excluded from ordinary Git history.