Novelty accounting
contributions
Each contribution is scored on the canonical four-tier novelty scale. Our self-claim ceiling is N3 (first-of-kind demonstration). N4 — paradigm-shifting / Nobel-worthy — is intentionally reserved for outside arbiters and is never self-claimed. Definitions, prior work, and verification links below.
How the research outputs fit together
The portfolio now follows three scientific questions, not a fixed paper count. Bounce theory contains the lead P2 calculation, a focused P1A boundary Note, and P1B verification software. DESI anomaly discovery is being rebuilt as a science-first flagship, with the current P3 retained as its public-ID provenance release. Galaxy chirality contains the lead P4 observed-label catalog/null result and the distinct P5 chirality–environment companion. These programs share methods and data, but they do not constitute a single evidentiary chain from bounce theory to a claimed survey detection.
Approved endpoint: six standalone works across the three programs, plus P3 as a supporting data release integrated with the rebuilt anomaly flagship. See the research programs and evidence library for full status and PDFs.
PARADIGM-SHIFTING (RESERVED)
Intentionally empty. N4 is reserved for paradigm-shifting, consensus-breaking, or Nobel-worthy results. That tier is awarded by the field over time — through broad community replication, citation, and consensus — not by the authors. We cap our own claims at N3 (first-of-kind demonstration / new constraint / new direction) and let outside arbiters raise the ceiling if any of this work earns it.
No BigBounce paper self-annotates at N4. Site copy, paper abstracts, and all contribution records are capped at N3 by internal review.
NOVEL COMBINATION / EXTENSION
Minimal-ECH Branch Clarification
A convention-audited account of what follows when the non-propagating connection is eliminated in the stated minimal Einstein--Cartan--Holst setup, including the spin-sourced contact term and the zero-spin scalar branch.
Why it matters
It supplies a bounded theoretical baseline for the bounce program: which minimal branches are addressed here, and which cosmological mechanisms are outside the Note's scope.
What it is
The Note consolidates standard identities, gives the contact-interaction coefficient and a scale benchmark, and separates spin-sourced from zero-spin scalar statements. It does not claim an all-orders observable theorem or a complete cosmological no-go result.
What existed before
Hehl et al. (1976); Freidel, Minic & Takeuchi (2005); Calcagni & Mercuri (2009); Mercuri (2009); de Berredo-Peixoto et al. (2012); Långvik et al.
What we did
The contribution is careful convention and scope control for this narrow branch; novelty and broader significance are for independent review.
How to verify
Exact Matter-Contraction Non-Gaussianity Derivation
A reproducible rederivation of the stated matter-contraction local amplitude, with explicit convention, bounce-transfer, and survey-mapping limits.
Why it matters
Makes a specific matter-contraction calculation inspectable and potentially testable. P2 rederives the squeezed value −35/16 and states the bounce-transmission and survey-mapping assumptions that must hold before it becomes an observational test.
What it is
A convention-pinned local non-Gaussian amplitude for the stated matter-contraction setup, paired with conditional sensitivity studies. The survey numbers are forecasting diagnostics, not a detection or a unique proof of a bounce.
What existed before
Cai, Xue, Brandenberger & Zhang (2009); Heinrich, Doré & Krause (2023); Dalal et al. (2008); Li & Brandenberger (2014).
What we did
A reproducible derivation plus explicit accounting of convention, template, transmission, projection, and survey-covariance boundaries. Novelty and venue significance remain for independent review.
How to verify
Exact Ordered Four-Vertex Polynomial
Re-sums the four stated cubic vertices in an explicitly ordered symmetric basis and makes the coefficient convention independently checkable.
Why it matters
It makes the calculation's ordered-vertex convention inspectable and separates that derivation from conditional observational forecasts.
What it is
The exact rational re-summation gives the unique ordered-basis coefficients (3, 1, -9, 5, -33, 9) and the squeezed amplitude −35/16. Independent checks use the order-grouped expressions and the general-c_s formula; the result is algebraic and scoped to the stated contraction-phase action.
How to verify
DESI Observed-Label Chirality Catalog and Dipole Null
An 8.47-million-row DESI observed-label catalog with a declared 890,069-object high-confidence sample and a primary dipole result consistent with zero.
Why it matters
It tests an observed-label chirality claim at scale while keeping label-transfer and training-composition limits explicit. It is not a physical primordial-parity constraint.
What it is
The release contains 8,474,531 DESI Legacy DR8 labels. Starting from 949,584 high-confidence rows, the declared safety quarantine removes 59,515, leaving 890,069 quality-controlled rows; 887,472 enter the supported-pixel fit. The primary result is null-consistent at z_mom=+0.635 with one-sided rank p=0.23768. Coverage-calibrated injection–recovery gives an observed-label sensitivity A95_obs≈0.98%, not a physical parity bound.
How to verify
Supporting DESI Public-ID Recovery (181 TARGETIDs)
A focused, reproducible public-ID recovery of a frozen historical DESI DR1 anomaly list: 181 warning-free global-primary TARGETIDs, split transparently into 170 high-coordinate-consistency core associations and 11 lower-confidence positional associations. This is an archive-recovery / provenance product — explicitly NOT a purity, novelty, or detection claim.
Why it matters
Reproducible archive provenance — recovering public IDs from a surviving historical anomaly list, with exact source-row lineage — is a transparency contribution: a reviewer can replay the exact selection waterfall and coordinate-association quality tiers. The paper deliberately declines physical classification, purity, novelty, and anomaly-rate claims that the surviving lineage does not support.
What it is
A declared 1-arcsec positional join over 20,299,155 eligible DESI rows → 2,468 positional parents → 2,448 global-primary rows → 181 warning-free associations (170 at ≤0.1″, 11 between 0.1″ and 1″). Every released row and all 18 carried DESI fields are re-read exactly from the recorded FITS row; the release carries exact source-row provenance, quality tiers, warned-row auxiliary data, shift controls, checksums, and a clean-checkout validator. As of r10 every claim site states that the sub-0.1″ core excess is expected seed self-recovery (the single-member cluster centroid equals the seed member's own coordinates by construction), not independent association evidence.
How to verify
Conditional SPHEREx Sensitivity Mapping
Conditional survey-sensitivity diagnostics for the stated matter-contraction amplitude; not a detection forecast guaranteed by P2.
Why it matters
Illustrates what additional bounce-transfer and survey assumptions would be needed to turn the derivation into an observational test.
What it is
The exact shape maps onto the published Heinrich et al. baseline with flat-grid recovery r=0.8354 (adopted r=0.84) and shape cosine 0.9817. The illustrative map is 2.63σ before additional nuisance marginalization; a channel-native surrogate spans 3.5σ with nuisances fixed, 3.1σ after A_GR marginalization, 2.3σ with a 30% b_φ prior, and 0.4σ when b_φ is free.
How to verify
DESI Chirality × Environment Null (z-shell corrected)
A catalog-native, exploratory chirality--environment comparison whose focal contrast is consistent with zero under declared controls and sensitivity checks.
Why it matters
It asks a distinct environment-dependence question using P4 labels; it is not preregistered, independent of P4, or a physical-handedness constraint.
What it is
The released DESIVAST GALZONE universe contains 694,642 unique TARGETIDs. Joining P4 yields 145,789 rows; 145,766 OUT=0 rows form the quality parent, split into 31,937 void and 113,829 non-void rows. The adjusted non-void-minus-void contrast is +0.00145442 with SE=0.00331502, 95% CI [−0.00504290,+0.00795174], normal p=0.66085, and wild-cluster p=0.67345.
How to verify
Archived MCMC Verification Infrastructure
Frozen posterior records document earlier verification work. They are retained for provenance and are not a current lead result; P1B is namaster-proof research software.
Why it matters
Demonstrates honest negative reporting — we found our own bug and disclosed it.
What it is
Cobaya/CAMB chain artifacts and diagnostics are preserved as historical research records. Their prior publication assignment and headline interpretations are superseded by the approved portfolio map.
How to verify
Archived NANOGrav Slope Comparison
A simplified historical power-law slope comparison is preserved for reproducibility; it does not identify a bounce origin or exclude complete astrophysical alternatives.
Why it matters
It records an exploratory model-comparison path and, more importantly, the boundary that a single fitted slope cannot determine physical origin.
What it is
The archived computation uses a public summary likelihood and a simplified slope family. It is not part of the selected lead claims and would require a preregistered, multi-model physical analysis to revisit.
namaster-proof Verification Software
An installable verification library for exact NaMaster bandpower windows and content-bound execution receipts; it is software infrastructure, not a sky measurement.
Why it matters
It turns two reproducibility hazards into testable software contracts that other pseudo-C_ℓ analyses can reuse.
What it is
The package exercises exact bandpower-window application, validates receipt/result byte binding, and ships tests and examples for the two failure modes described by P1B. Historical birefringence and chirality experiments remain program provenance and are not P1B headline results.
Provenance-Audit Methodology (retract-and-rebuild)
Artifact-level provenance checks caught unsupported intermediate claims, which were withdrawn before submission and replaced only when a current, reproducible result existed.
Why it matters
Most published nulls and detections never face an artifact-level audit. Treating retraction-and-rebuild as a first-class, documented workflow is itself a transparency contribution — reviewers can replay the exact decision chain.
What it is
The workflow binds claims to file hashes, generators, footprint checks, exact inputs, and rerunnable outputs. Failed provenance does not become a caveat attached to a headline; it removes the headline until a supported replacement exists. Current paper values must come from each final candidate, not superseded audit-era numbers.
12-Job Compute Reproducibility Chain
All 12 load-bearing compute closures (fsky sweeps, continuous-prior MCMC, permutation rebuilds, sign-symmetry reruns) executed on a dedicated pod with committed scripts + JSON artifacts, for ~$0.55 total.
Why it matters
A reviewer can rerun any headline number from the committed chain — reproducibility as an artifact, not a promise.
What it is
Each archived job ships its driver, inputs, seeds, and output artifact. Earlier P1B/MCMC assignments are superseded; the current P1B contribution is namaster-proof. P4/P5 receipts remain supporting reproducibility evidence for their scoped claims.
How to verify
Archived BigAE Multi-Survey Pipeline
Historical autoencoder pipeline artifacts are preserved, but their cross-survey counts and validation claims are not the current anomaly flagship and are not acceptance targets for the clean DESI rerun.
Why it matters
The archive records useful lessons about domain shift, native retraining, and provenance. Its scientific claims must be regenerated under the new fail-closed model/input/scaler contract.
What it is
Legacy per-survey retrains exposed severe cross-transfer domain shift. The new flagship must regenerate the DESI sample with immutable inputs, the hash-bound model, a sealed scaler, shard receipts, deterministic deduplication, and independent candidate validation.
What existed before
Autoencoder outlier detection (Baron & Poznanski 2017); single-survey spectral anomaly searches.
What we did
No current novelty claim is assigned to the unreconciled multi-survey catalog. The reusable contribution is the preserved failure analysis and the clean-rerun contract now governing the rebuild.
How to verify
Z₂-Flip-Equivariant Chirality Classifier (released model)
A released classifier/checkpoint used to produce observed chirality labels, with reflection-aware processing and documented training-composition limits.
Why it matters
Reflection-aware processing is a useful control for orientation-sensitive labels, but it does not calibrate true spin or remove every survey and training systematic.
What it is
The released ViT-based checkpoint supports P4/P5 reproducibility. P4 documents unresolved historical training-composition conflicts and an unreproduced CE-included accuracy path; the released catalog labels remain observed classifier outputs, not calibrated physical handedness.
What existed before
CNN/ViT galaxy-morphology classifiers (Galaxy Zoo DECaLS, Zoobot); standard non-equivariant chirality classifiers (Shamir et al.).
What we did
The reusable output is the released checkpoint plus its declared scope and provenance boundary; priority claims are left to independent review.
How to verify
Historical GZ1-Only Control
A lower-power human-label control is preserved as supporting evidence; it neither proves a physical null nor replaces P4's current quality-controlled primary estimator.
Why it matters
The control probes sensitivity to one training path while leaving morphology transfer, footprint, and upstream imaging systematics unresolved.
What it is
The archived GZ1-only run is null-consistent at lower power. It is a historical robustness control, not the current P4 headline and not a physical-parity constraint.
What existed before
Self-training / pseudo-label validation typically checks classifier accuracy, not downstream-measurement independence.
What we did
Useful supporting control with a deliberately bounded interpretation.
How to verify
INCREMENTAL REFINEMENT / REPLICATION
Archived ECH Route Map
Earlier route-mapping material is retained as provenance, not presented as a current universal closure result or as the scope of P1A.
Why it matters
It documents how the program narrowed; it is not a selected scientific claim.
What it is
The historical map contains exploratory constraints and branch notes. It must not be read as closing every minimal-ECH route to dark energy or as a replacement for a model-specific analysis.
What existed before
Blagojević & Hehl (2013); Weinberg (1989); 't Hooft (1979); Shie, Nester & Yo (2008).
What we did
Preserved for traceability only; no current novelty or closure claim is assigned.
How to verify
Archived ALP Birefringence Exploration
Earlier exploratory ALP calculations are retained as provenance; they are not a selected claim of P1A or P2.
Why it matters
It records an exploratory line without implying a present ECH prediction or measurement.
What it is
The archive contains exploratory ALP evolution and synthetic-estimator checks. None is assigned to the selected P1A or P2 scientific claims, and the material is not presented as a sky measurement or an ECH prediction.
How to verify
Archived Nieh--Yan Branch Note
An earlier conceptual branch is preserved for traceability; it is not a general theorem or a current P1A conclusion.
Why it matters
The item explains prior exploration, not the present publication claim.
What it is
The archive records a conditional discussion of topological and non-topological Nieh--Yan constructions. It does not establish universal mutual exclusivity across gravitational models.