V12 Ensemble Discovery Scan: Unseen Euclid Q1 Strong Lens Candidates

Inference results from the V12 ensemble model (V7×0.3 + V10×0.2 + V11×0.5, weighted logit averaging, AUROC=0.8871 on sacred test set) applied to 1,415 unseen Euclid Q1 ERO images — all files outside the locked training, validation, and test split. Contamination verified at zero overlap via skycoord crossmatch <2 arcsec against all 1,000 sacred split files.

Scanned data: 247 Grade B probable-lens candidates + 1,168 Grade C holdout files from the ESA SLDE Q1 catalog (VIS band, 300×300px FITS cutouts, 0.1 arcsec/px). Inference used 4-angle TTA (0°/90°/180°/270° rotations, averaged softmax). Classification threshold: P≥0.70.

Key results:- 33/86 pure Grade B candidates (no NEG prefix) flagged as lens at P≥0.70 (38.4%)- 228 files with all three component models independently scoring P>0.90- 625 total files flagged; 790 not flagged- Grade C false positive rate reflects documented out-of-distribution   generalization limit consistent with ESA arXiv:2512.05899

Package contents:- v12_unseen_q1_scan_full.csv: All 1,415 files with per-model TTA and   no-TTA scores (p_v7, p_v10, p_v11, p_ensemble), RA, Dec, grade,   classification- v12_unseen_lens_candidates.csv: 625 flagged candidates only (P≥0.70)- v12_unseen_q1_scan_full.xlsx: Same as CSV in Excel format- v12_discovery_gallery_final.pdf: 32-page visual gallery of top-ranked   candidates, 4 per page, sorted by ensemble TTA score descending,   annotated with individual model scores

These are lens candidates for follow-up visual inspection, not confirmed lens discoveries. Real lens prevalence in photometric surveys is ~0.1–1%.