Galaxy Chirality Classifier (bamfai/galaxy-chirality-v2)

Companion model for Paper IV v1.0.128 — A Survey-Scale Chirality Catalog of 8.47M Galaxies (3.2M Spirals): A Null Detection of Large-Scale Parity Violation at Sub-Percent Sensitivity.

Source: Hubify-Projects/bigbounce @ tag paper4-v1.0.128.

Model overview

• Architecture: ViT-Small/16 with mirror-equivariant test-time augmentation. Z₂ 2-fold flip is the production TTA mode (D₄ 8-fold tested on hold-outs only; see Paper IV §sec:tta).
• Output classes: CW, CCW, NS (not-spiral).
• Calibration: per-class Platt scaling (L-BFGS on a held-out 20% split). Raw → calibrated → equivariant residuals reported in Paper IV Table V (raw +0.79% / 28.8σ → calibrated +0.4% / 14.6σ → equivariant −0.26% / 9.5σ).
• Equivariance suppression factor 3.86× (raw asymmetry +2.05% → equivariant asymmetry −0.53%).

Catalog scale

Applied to the DESI Legacy DR9 8.47M-galaxy footprint via the canonical Paper IV pipeline:

• 8,474,531 galaxies classified (1,687,069 CW / 1,634,726 CCW / 5,152,736 NS)
• 3,201,160 chirality-relevant spirals
• Catalog-wide CW fraction (post-TTA equivariant) 0.4974 ± 0.000279, consistent with parity at ~1σ; the residual −0.0026 monopole is a classifier-residual bias, not a cosmological dipole.

Key results (v1.0.128)

• −0.12σ MASTER-deconvolved ℓ=1 amplitude on the subsample-mask (load-bearing null).
• +3.64σ canonical-mask ℓ=1 amplitude (post-MASTER) interpreted as a coherent depth/PSF/morphology-correlated systematic on the canonical footprint, NOT a primordial dipole, via three direct quantitative anchors:
  • ℓ=2 cross-spectrum quadrupole anti-alignment with pixel-density proxy (r = −0.65, σ = −2.89)
  • Leg-stratified ℓ=1 contribution: BASS+MzLS / DECaLS / DES cross-power gives induced |a₁| = 25% of canonical |a₁_obs| = 7.04×10⁻³
  • MASTER-decoupled monopole-only null × 500: data post-MASTER C₁ = 6.55×10⁻⁶ vs null mean 8.0×10⁻⁷ (σ = +4.84 Gaussian / empirical-rank p = 0.006); monopole leakage accounts for ~12% of post-MASTER C₁
• Family-level max-stat null on 15-cell leg×conf grid: observed max|σ| = 4.72 collapses to ~2.4σ family-corrected joint p = 0.0086.

v1.0.128 changelog (Houston-shared 3-reviewer external review + 2 follow-up MAJOR closures)

• v1.0.123 closed the Houston-shared external review on v1.0.122 (ChatGPT MAJOR REVISIONS + Grok MINOR REVISIONS + Gemini MAJOR REVISIONS; 9 BLOCKER + 10 MAJOR bundled hard-fix; 2 BLOCKERs audit-falsified). Stripped all internal-review scaffolding (Perplexity R22, P4-EXT, P4-INT, R20-Grok-B2 closure, etc.); fixed 3 undefined section references; tightened Shamir 2022 framing; softened "three-interpretation closure" → "diagnostic"; added explicit "headline uses p_CW-weighted maps not argmax labels" defense against the 21% D4-TTA flip-rate concern. HF dataset README also pushed to v1.0.123 in the same fire.
• v1.0.124 closed ChatGPT MAJOR M3: new mask pixel-count-threshold robustness sweep (5 cells × MASTER coupling matrix × N=200 monopole-only null) shows the canonical-mask ℓ=1 σ is robust at +6.31 to +8.26 across n_total > {1,5,10,20,50}. Signal does not attenuate at higher thresholds — rules out the low-count-edge artifact interpretation. New §IX.B "Mask robustness" subsection
  • Table V + pre-specified estimator hierarchy paragraph.
• v1.0.128 closed Gemini MAJOR M-2: new DECaLS [0.5,0.6) stratum-specific cross-spectrum C^{An}_ell on the 938,563-spiral stratum (f_sky=0.279). Result: r_ℓ=1 = −0.70 (σ = −1.68), same negative-correlation sign as full canonical −0.49 with larger magnitude. The DECaLS-stratum excess is anti-correlated with its own pixel-density at the same scale as the canonical depth-systematic signature — directly ties the stratum's family-corrected ~2.4σ excess to the depth-correlated systematic family (interpretation ii), not a separate DECaLS-specific physical signal.

R-round cross-vendor convergence

Cascaded-loop exit confirmed at Paper IV v1.0.128 per AGENT_RULES §4.4.1: 3 consecutive 5/5 clean R-rounds (R23 + R24 + R25) across DeepSeek-V4-Pro

• Gemini-3.1-Pro + GPT-5 + Grok-4.3 + Perplexity-Sonar-Pro (45 of 45 reviewers returned 0 BLOCKER / 0 MAJOR).

Usage

from huggingface_hub import hf_hub_download
import torch

ckpt_path = hf_hub_download(
    repo_id="bamfai/galaxy-chirality-v2",
    filename="model.safetensors",
    revision="paper4-v1.0.128",
)
state = torch.load(ckpt_path, map_location="cpu", weights_only=True)
# Then load into ViT-Small/16 + 3-class head as documented in
# pipelines/p2_chirality/run_eq_dataloader.py

Companion catalog dataset

Per-galaxy CW/CCW/NS labels + per-class probabilities + leg provenance: bamfai/galaxy-chirality-catalog (immutable revision paper4-v1.0.128).

Citation

@misc{golden_chirality_2026,
  author = {Houston Golden},
  title  = {A Survey-Scale Chirality Catalog of 8.47M Galaxies (3.2M Spirals):
            A Null Detection of Large-Scale Parity Violation at Sub-Percent Sensitivity},
  year   = {2026},
  url    = {https://github.com/Hubify-Projects/bigbounce/blob/main/pipelines/p2_chirality/chirality_catalog_paper.tex},
  note   = {Paper IV, version v1.0.128; tag paper4-v1.0.128}
}

Bias and limitations

• Trained on DESI Legacy DR9 / DR8 imaging. Generalization to other surveys (HSC, KiDS, future LSST) requires re-validation.
• A residual catalog-wide CW-fraction offset of −0.0026 (a ~5σ deviation from 0.5 at N = 3.2M chirality-relevant spirals) is attributable to classifier-residual bias and is documented in Paper IV §VI.E.
• For environment-dependent chirality, see the companion Paper V (pipelines/p5_desi_chirality/paper/p5_desi_chirality.tex).

Provenance

• Released: 2026-05-21 (cron fire #3, autonomous drive-to-100 loop).
• Paper IV tag: paper4-v1.0.128.
• Repository SHA: see git log paper4-v1.0.128 in Hubify-Projects/bigbounce.
• Prior model card revision: v1.0.104 (superseded).