JasonResearch/inference-lens-demo: Inference Lens Demo v1.0

Inference Lens Demo: Producing Dark-Matter-Like Discrepancies Without Adding Mass This archive provides a controlled, reproducible demonstration that a dark-matter-like dynamical discrepancy can arise purely from measurement and inference differences, even when the underlying (“true”) galaxy contains no dark matter. The demonstration is intentionally simple and didactic. A fixed baryon-only galaxy is held constant while two distinct observational / inference pipelines (“Lens A” and “Lens B”) are applied. The true mass distribution is unchanged in all cases. Only the observational depth, resolution, and inference assumptions differ. The resulting inferred dynamical discrepancy changes accordingly. This archive does not claim that dark matter is an optical illusion or that real galaxies lack non-baryonic mass. Instead, it provides an existence proof: inference and normalization choices alone are sufficient to manufacture a discrepancy signal that resembles dark-matter phenomenology. Demonstration Concept The demo enforces three constraints: The true baryonic mass distribution is fixed and known. Two observational / inference pipelines are applied to the same system. Any change in inferred dynamics must therefore arise from inference, not physics. Under these conditions, the demo shows that: observable extent can differ between pipelines, surface-brightness truncation and information loss alter inferred structure, and standard discrepancy proxies (e.g., Mdyn/MbM_{\mathrm{dyn}}/M_bMdyn/Mb, J3J_3J3) can diverge without adding mass. Outputs A single run produces: A three-panel figure showing: the true rotation curve vs. inferred curves (Lens A vs. Lens B), observed surface-brightness profiles and truncation behavior, a comparison of discrepancy proxies (e.g., log⁡10(J3)\log_{10}(J_3)log10(J3) or Mdyn/MbM_{\mathrm{dyn}}/M_bMdyn/Mb). A CSV summary containing truth values, inferred quantities, and mismatch metrics for both pipelines. A saved configuration file documenting the exact parameters used for the run. Reproducibility All runs can be made deterministic using a fixed random seed. Each output directory contains a complete snapshot of the configuration used, enabling exact reproduction of any result. The code is intentionally lightweight, uses standard scientific Python packages, and can be audited end-to-end in a short time. Scientific Context This demonstration is an illustrative companion to the broader Observer-Normalized Scale Relativity (ONSR) research program, which investigates how observer-centric normalization and inference choices affect global physical interpretation. Related archived works include: Observer-Normalized Scale Relativity: A Coherence Constraint on Global Physical Inference (core framework)https://doi.org/10.5281/zenodo.18356675 Stress Testing Observer-Normalized Coherence in Disk Galaxieshttps://doi.org/10.5281/zenodo.18294837 Predictive Results in the Dwarf-Galaxy Regimehttps://doi.org/10.5281/zenodo.18308519 Those works address empirical and theoretical consequences in real galaxy data. This archive isolates the mechanism in a minimal, transparent setting. Scope and Non-Claims This archive does not: fit galaxy data, modify gravitational laws, introduce new forces or particles, claim that dark matter is fictitious, replace empirical cosmological models. It demonstrates how inference alone can generate dark-matter-like signals under realistic observational limitations. Execution Install dependencies: pip install -r requirements.txt Run the default A/B demonstration: python run_demo.py --out outputs/demo --seed 1 --save-plots Run individual presets: python run_demo.py --preset presets/lens_A_deep_sharp.json --out outputs/runA --seed 1 --save-plots python run_demo.py --preset presets/lens_B_shallow_blurry.json --out outputs/runB --seed 1 --save-plots Archival Note This archive is intended as a conceptual and pedagogical artifact. It demonstrates that dynamical discrepancies can arise from inference structure alone, establishing a concrete baseline for interpreting empirical results in observer-normalized frameworks.