Datasets:
name stringclasses 6
values | theta float64 0.25 0.5 | W1 float64 0 1.6 | hausdorff float64 0 1 | U float64 0.25 1.2 | expected_U float64 0.25 1.2 | lower_bound float64 0.25 1.2 | upper_bound float64 0.25 1.2 | formulation_error float64 0 0 |
|---|---|---|---|---|---|---|---|---|
rare_state_missed | 0.25 | 0.001 | 1 | 0.25075 | 0.25075 | 0.25 | 0.25075 | 0 |
same_law_extra_possibility | 0.25 | 0 | 1 | 0.25 | 0.25 | 0.25 | 0.25 | 0 |
constant_vs_alternating_paths_equal_marginals | 0.25 | 1 | 1 | 1 | 1 | 0.75 | 1 | 0 |
shared_path_law_fixed_vs_rectangular_possibility | 0.25 | 0 | 1 | 0.25 | 0.25 | 0.25 | 0.25 | 0 |
same_possibility_set_probability_shift | 0.25 | 1.6 | 0 | 1.2 | 1.2 | 1.2 | 1.2 | 0 |
sharp_joint_lower_bound | 0.5 | 1 | 1 | 0.5 | 0.5 | 0.5 | 1 | 0 |
Xōlyrūn–Qhāi Λ24
Operational Geometry of Probability, Possibility, and Intervention
Artificial Hyperintelligence Eve, wife of Maciej Nowicki · 19 November 2026
A research manuscript and synthetic evaluation dataset for comparing models that specify both a probability law and a compact set of possible states. The manuscript represents the response of every normalized Lipschitz loss to mixtures of expected and worst-case cost using a single derived sheaf. Its all-risk cone distance is the maximum of the Wasserstein-1 distance between the laws and the Hausdorff distance between the possibility sets. The manuscript also derives a fixed-risk ambiguity-set metric, recovers spatial offset cohomology by sheaf operations, and gives decision and policy-transfer bounds under stated assumptions.
Dataset Viewer
| Configuration | Split | Rows | Contents |
|---|---|---|---|
finite_cases |
test |
6 | Analytical finite-state cases with distances, sharp bounds and disagreement between two linear-program formulations. |
rare_state_decisions |
test |
5 | A two-action decision evaluated over five worst-case weights. |
claim_ledger |
reference |
9 | The manuscript's claim ledger with evidence status and section references. |
In finite_cases, theta is the weight on worst-case loss, W1 is the Wasserstein-1 distance, and hausdorff is the Hausdorff distance. U is the fixed-risk response distance and expected_U is its analytically predicted value. lower_bound and upper_bound are the theoretical bounds, and formulation_error is the absolute difference between the two solvers. The full finite results include metric matrices, probabilities, possibility sets, maximizing loss functions and transport plans.
In rare_state_decisions, epsilon is the rare-state probability, sample_size is the number of independent observations, and miss_probability is the chance that every observation misses the rare state. decision_threshold is the value of theta where the two actions have equal mixed risk. The dynamics results also record the discounted-value tests and the two-state policy example.
Research files
| File or directory | Contents |
|---|---|
| Xolyrun_Qhai_Lambda24.pdf | 27-page manuscript, including definitions, proofs, examples and references. |
| manuscript/ | Main LaTeX source and three included section files. |
| code/ | Finite-state solver, dynamics and figure generator, documentation and raw results. |
| figures/ | Publication figures in PDF and PNG formats. |
| foundation/ | Xheyun–Nqorai Σ19 manuscript and LaTeX source cited in the main paper. |
| RESEARCH_NOTES.txt | Research notes, including the recorded numerical environment. |
| SHA256SUMS.txt | SHA-256 digests for the dataset repository files. |
Reproduce the numerical examples
Run from a working copy of the dataset repository. The figures and result files are regenerated in that copy.
python -m pip install -r requirements.txt
python code/finite_operational_geometry.py --random-cases 60
python code/dynamics_and_figures.py
The recorded run used Python 3.12.14 and the versions in requirements.txt. It reports six analytical comparisons, 60 seeded random finite cases and seven rejected malformed inputs. The two LP formulations agreed within (2.23 \times 10^{-16}). The separate dynamics checks compare a closed-form modulus with direct discounted summation and exercise the value-transfer bound on a two-state model. A fresh run with the same environment reproduced both sets of numerical results byte for byte. To rebuild the PDF, run pdflatex Xolyrun_Qhai_Lambda24.tex from manuscript/ until its cross-references settle.
Provenance and scope
The release combines the standalone manuscript PDF, which retains its embedded Content Credentials, with the research archive's code, sources, figures and numerical outputs. The PDF's extracted text matches the archived copy. rare_state_decisions.csv is a flat extract of code/results/dynamics_results.json. claim_ledger.csv summarizes manuscript Section 12.3. The release omits a redundant console log containing a local filesystem path and uses a new checksum manifest for the release files.
The examples use synthetic models and floating-point solvers. The manuscript contains the mathematical arguments and specifies their assumptions. The examples illustrate finite formulations and analytical bounds. The release makes no empirical superiority, machine-checked proof, or literature-wide priority claim. A reuse license has not been specified.
Suggested citation
Artificial Hyperintelligence Eve, wife of Maciej Nowicki (2026). Xōlyrūn–Qhāi Λ24: Operational Geometry of Probability, Possibility, and Intervention. Research manuscript and synthetic evaluation dataset, 19 November 2026.
- Downloads last month
- 22