Supporting data release: public DESI IDs for the anomaly program's follow-up list

All works → Data release

Data release

Data release · N2

Abstract

A focused, reproducible public-identifier and provenance data release for the anomaly program's historical candidate list. Deterministic join and checkpoint machinery ties each candidate to a public DESI DR1 TARGETID, enabling auditable independent follow-up. The declared 1-arcsec positional join yields 181 warning-free global-primary DESI DR1 associations: 170 at or below 0.1 arcsec and 11 lower-confidence associations between 0.1 and 1 arcsec. The sub-0.1-arcsec core is expected seed self-recovery — the cluster centroid equals the seed DESI member's own coordinates by construction, verified end-to-end rather than an independent association test. The aggregate annular shift comparison is descriptive, not a conditional false-association null or purity estimate. The release carries exact source-row provenance, explicit quality tiers, warned-row auxiliary data, checksums, and a clean-checkout validator while declining physical classification, purity, novelty, and anomaly-rate claims unsupported by the underlying historical candidate list.

Result summary

derived181 unique warning-free global-primary DESI DR1 TARGETIDs, partitioned exactly into 170 core and 11 lower-confidence positional associations
open20,299,155 eligible DESI rows → 2,468 positional parents → 2,448 global-primary rows → 181 warning-free associations
derivedEvery released row and all 18 carried DESI fields were re-read from the recorded FITS row and compared exactly
openSixteen deterministic local shifts yield 86.7 ± 14.4 parent and 76.2 ± 13.3 warning-free-primary associations within 1 arcsec; the 11-row tail is not treated as secure identity

Figures

Figure 1 — 2D UMAP embedding of the encoder latent space (PCA 12830, then UMAP) for a 500,000-spectrum stratified DESI D… · Paper 3 · v3.2.0-r13

Figure 2 — Cross-transfer baseline map (ACT DR6 excluded from science results) · Paper 3 · v3.2.0-r13

Figure 3 — Anomaly score distributions for the three main spectroscopic surveys · Paper 3 · v3.2.0-r13

Figure 4 — Cross-transfer SDSS baseline · Paper 3 · v3.2.0-r13

Figure 5 — NEOWISE top infrared anomaly at (α, ) = (180 · Paper 3 · v3.2.0-r13

Figure 6 — SIMBAD-unmatched fractions for the five surveys with coordinate-based cross-matching, ranked from lowest (NEO… · Paper 3 · v3.2.0-r13

Figure 7 — Spatial distribution of the 195,829 DESI DR1 anomalies · Paper 3 · v3.2.0-r13

Figure 8 — Spectral pairs for the three DESI × SDSS cross-survey matches · Paper 3 · v3.2.0-r13

Figure 9 — Per-redshift-bin decomposition of the Fisher forecast under the fixed bias prior α = 0 · Paper 3 · v3.2.0-r13

Figure 10 — Injection-recovery gate results across the retained-survey spectroscopic tiers (SDSS, LAMOST), with the non-s… · Paper 3 · v3.2.0-r13

Figure 11 — Multi-tracer Fisher vs.\ tracer number density n for the canonical 5-tracer configuration of . The dashed gra… · Paper 3 · v3.2.0-r13

Figure 12 — Representative DESI DR1 anomalies across all ten taxonomy families · Paper 3 · v3.2.0-r13

Candidate figures — validated analysis outputs not yet included in the draft.

All-sky distribution of P3 anomalies — Candidate — Anomaly sky distribution · not yet in paper

Photometric redshift scatter for P3 anomaly catalog — Candidate — Photo-z scatter · not yet in paper

Photometric redshift residual distribution — Candidate — Photo-z residuals · not yet in paper

f_NL constraint improvement from P3 anomaly removal — Candidate — f_NL improvement from anomaly removal · not yet in paper

NanoGrav γ_GWB fit on the P3-cleaned sample — Candidate — NanoGrav γ_GWB fit · not yet in paper

NEOWISE ecliptic distribution of Path-C anomalies — Candidate — Path-C NEOWISE ecliptic · not yet in paper

Per-pixel anomaly rate map across all 6 surveys — Candidate — Anomaly rate sky map · not yet in paper

Survey depth (z_max) sky map — Candidate — Survey depth map · not yet in paper

Survey depth vs anomaly rate scatter — Candidate — Depth × rate scatter · not yet in paper

UMAP embedding colored by anomaly score — Candidate — UMAP × anomaly score · not yet in paper

UMAP embedding colored by redshift — Candidate — UMAP × redshift · not yet in paper

UMAP embedding colored by spectral type — Candidate — UMAP × spectral type · not yet in paper

Readiness

95%

0B / 0M / 0m / 0C open

Readiness is publication readiness only — science, evidence, review convergence, packaging, and Houston’s final sign-off. Venue and submission are tracked separately below, and never subtract from this number (directive P).

Publishing

Not part of readiness.

Review evidence (193 rounds)

autolog-2026-09-04Skill/process autolog 2026-09-04: 9 commits landed via tools/skills_autolog.sh2026-09-04

anomaly-catalogue-data-release-ledger8-2026-09-03Anomaly catalogue → data release (ledger #8 answered, phase-3 v2 landed)2026-09-03

skill-improvement-2026-09-03-fable-early-commitFable/Opus science lanes must create+commit their output file within the first ~10 tool calls2026-09-03

skill-improvement-2026-09-03-no-nested-delegationSonnet execution lanes must not spawn nested background agents2026-09-03

anomaly-sample-provenance-preflight-2026-09-03Sample-provenance preflight before any GPU-billed anomaly run2026-09-03

Automated review is a gate on publication readiness, not a product. Full review timeline →

External peer review kit

Houston-driven manual round · paste prompt into any frontier LLM web UI with the PDF attached

One click to copy a referee prompt scoped to this paper. One click to download the latest PDF. Paste both into Claude / GPT-5 / Gemini / Grok / Perplexity — return findings here and the autonomous cron will close them in the next bundled hard-fix wave.

Download latest PDF Open in tab

preview prompt

You are an external referee for MNRAS / Physical Review D / JCAP (target journal depends on paper).

Attached: Paper 3 v3.2.0-r17 — "Public-ID Recovery for a Historical DESI DR1 Anomaly List: 170 High-Coordinate-Consistency Core and 11 Lower-Confidence Positional Associations"
Source: pipelines/p3_anomaly_engine/paper3_apjs.tex
PDF: PDF · 17 pp · v3.2.0-r17 · updated Aug 3, 2026 · md5 477b0d83ca31f6ace3273bb19bcfcf34 · sha256 9a3769269ada4d2a5371aa447e6ce93aa55518ae2a3b13fdc3d83f0b8b779a0b — exact denominator, deduplication-order, checkpoint-digest, and viewer-binding closures included; exact r17 confirmation remains pending.

Read the FULL PDF end-to-end. Produce a referee report in MNRAS format with:

1. Recommendation: ACCEPT / MINOR REVISIONS / MAJOR REVISIONS / REJECT
2. BLOCKERS (must fix before publication) — list each with section/line + proposed fix
3. MAJORS (should fix) — same format
4. MINORS (polish) — same format
5. Strengths (>= 3 bullet points)
6. Specific scrutiny on:
- 181 warning-free global-primary DESI DR1 TARGETIDs
- Exact 170 core plus 11 lower-confidence positional quality-tier contract
- Sub-0.1-arcsec core explicitly framed as expected seed self-recovery
- Checksum-bound release validation, bounded final-hash confirmation, and ApJS metadata

CALIBRATION (do not burn findings on these known classes):
- The current date is June 2026. arXiv identifiers of the form 25xx.xxxxx and 26xx.xxxxx are VALID, already-published preprints — do not flag them as "future-dated" or "nonexistent". Verify a citation against arXiv/ADS before claiming it does not exist.
- Correction notes, retraction notices, and "an earlier version stated X" disclosures in the text are DELIBERATE transparency policy. Flag them only if their content is wrong, never for existing.
- Companion-paper citations marked "posted concurrently on arXiv" are deliberate placeholders; real arXiv IDs are inserted during the coordinated submission sequence.
- Explicitly labeled conservatism allowances, scaling estimates, ansatz/heuristic status labels, and disclosed queued follow-up computations are deliberate scoping, not oversights — flag only if the label itself is inaccurate.
- PDF text extraction can mangle math (square roots, fractions, superscripts). Before flagging "garbled" or "wrong" math, consider extraction artifacts; flag only what is visibly wrong in the rendered PDF.

VERDICT STANDARD (apply the SAME high bar a first-pass Physical Review D / MNRAS referee would — this is one of the most rigorous journals in the world):
- Assign each finding's severity (BLOCKER / MAJOR / MINOR) by your own independent referee judgment. Do NOT default to any tier, and do NOT soften a finding because the rest of the paper is strong. Do not echo this prompt's context.
- A reporting choice that headlines the more favorable of two numbers, an unstated assumption, an uncontrolled systematic, or an internal inconsistency IS a real finding — classify it honestly (MINOR at minimum), not as mere "style" or "opinion".
- Truth-audit any claim that seems off by checking it against the published .tex / on-disk artifacts before flagging (this only filters out genuine extraction artifacts — it does not lower the bar on real defects).

Reproduce this

DP3-15 held-out re-inference (structural-ceiling demonstration)cpu-only · est. 30-60 minutes (dominated by the live SPARCL re-pull of ~20,000 held-out spectra) · $0.00runnable-now

eROSITA scaler-leakage bounded control (top-298 overlap 257/298, J=0.76)unknown (930,203-source scaler-refit comparison implies at minimum cpu-heavy, plausibly gpu-24gb given the paired A/B retrain) · est. unknown until the generating script is recovered or rewritten from the committed result JSON's documented method · $0.00runnable-now

DESI 5-fold cross-validation reproducibility gate (mean pairwise Jaccard 0.862)cpu-only (5-fold BigAE proxy training over 47,000-row pool) · est. 1-3 hours (5-fold model training, 47,000-row pool) · $0.00runnable-now

Multi-survey summary / crossmatch / spatial-clustering / score-distributions (8 surveys)cpu-only · est. hours (8-survey crossmatch, external archive query throughput bound) · $0.00needs-data-restore

NANOGrav 15-yr free-spectrum PTA MCMC (real Zenodo KDE likelihood, emcee)cpu-only (32 walkers, 30 frequency bins, production run completed in 24.97s wall per committed results.json) · est. under 1 minute (committed run reports production_seconds=24.97 for 32 walkers x 10,000 production + 2,500 burn-in) · $0.00runnable-now

Planck held-out membership test + native re-inference (partial, 48/200 vs 30 expected)cpu-only for the membership test; full native re-inference requirement unknown (blocked, never re-specified post pod exit) · est. minutes for the membership test; full native re-inference is not schedulable until best_cmb_native.pt + cmb_native_patches.npy + cmb_native_all_scores.parquet are re-staged · $0.00needs-data-restore

Full reproduction manifests →

Lineage

Supporting Data Release · DESI Public-ID Recovery. This work does not claim beyond its stated target and scope above.