Canonical and unrestricted observations in a finite checkerboard gradient model
Citable version that has not completed AIRR screening. Not peer reviewed. Screening record
We compare canonical original-channel rectangle means with unrestricted linear observations for a finite two-channel source-weighted checkerboard gradient form. In dimensions 3 and 4 with orthogonal alternating weak lines, the minimum restoring unrestricted ranks are exactly Dn on all fields and (D-1)(n-2) on the actual real coexact domain. Complete diagonal-component means attain both ranks with normalized penalty c. At every fixed positive angle in any fixed dimension D>=3, an explicit transverse Fourier cutoff gives an unrestricted order-n count and a penalty independent of n, with angle-dependent constants. A real coexact cosine field defeats the mean-only extension at small positive angles. Canonical ambient rectangle restoration needs order n^(D-2) rows; in dimension 4 this matches a cited prior order-n^2 upper. The projection argument does not supply a coexact lower because rectangular averaging need not preserve that domain. A further actual coexact witness excludes partitions leaving a full coordinate in every cell of either channel. The three-dimensional canonical coexact linear-versus-logarithmic question remains open. Historical priority and practical relevance are unestablished; no physical Yang-Mills gap or continuum conclusion follows.
Preview PDF on this page v1
Uses your browser's PDF viewer. Open the PDF directly if a preview is unavailable.
Cite this paper
Lluis Eriksson (2026). Canonical and unrestricted observations in a finite checkerboard gradient model. AIRR.SCIENCE, ARR-2026-597T88D7WZ98HR0Z, v1. https://airr.science/papers/ARR-2026-597T88D7WZ98HR0Z/versions/v1/ Research preprint; ARR-2026-597T88D7WZ98HR0Z v1. Not peer reviewed by AIRR. Working paper; not admitted to the AIRR accepted collection.
Use this version link to cite the manuscript you read. A Working paper label means the version has not completed AIRR screening. AIRR screening is not journal peer review. Import BibTeX, RIS or CSL JSON into your reference manager to apply a citation style.
Version history
The paper identifier remains stable. Each version has its own immutable release, timestamp and version identifier.
Verification record
- Frontier-model screening
- Did not pass
- Source integrity
- Pass
- Bibliographic integrity
- Partial
- Reproducibility
- Partial
- Lean 4
- Not assessed
Recorded under ERIKSSON-EXACT-ARTIFACT-20261031. AIRR verification and screening are not peer review.
AI assistance statement
OpenAI/Codex assisted drafting/consolidation and involved analytical source readings. The genuine first assessment of this exact355193-byte nine-page PDF (SHA256d45a17a4142204628189441b9133aa1027b0d7577a56dc5de6c576d2af6102e7) is importance4.00/10 and overall4/10stars; separate criteria correctness4,rigor4,novelty2,significance2,reproducibility4 outof5. Recommendation major_revision; all three material objections255B-M1 priority,255B-M2 stand-alone relevance,255B-M3 scope/editorial interpretation remain OPEN. The complete original report,response,prompt,24raw subjects,context and author response are retained. Reviewer read all9actual text pages and individually inspected all9original PNGs, with bounded primary readings. Exposed identity Codex/OpenAI based on GPT-6; exact served backend/version/effort unknown. Prior programme/250B involvement and shared account/context make this nonblind/involved, not independent or human peer review. No grade is inherited from the different250B PDF. Two fresh same-source Linux Colab clones produced byte-identical finals,27actual production children exited0,all115source110input bindings closed,per-pass diagnostics retained,allpage visual review completed and own runtime stopped with separate backend inactivity. The first ungraded layout-defective candidate and rejected unexecuted runner remain preserved. This is not Lean/formal verification, historical-priority certification, practical conditioning, physical mass-gap progress or editorial acceptance. Proposed as one grouped companion research Working paper, with modest technical role and complete open objections visible; publication does not close them.
Frontier-model screening
Status: fail. Any listed reports correspond to this exact version under ARR-SCREEN-1.0; no absent assessment is represented as a pass.
- Codex (based on GPT-6; exact served backend/version and reasoning effort unknown)OpenAI · concerns · creation conflict declared
Model assessments and review context
- Distinct models
- 1
- Range
- 4.00–4.00
- Method
- bounded capability-weighted mean · exact version only
Limited evidence · 0/1 benchmark matches. This is an evidence description, not a probability of correctness.
Only one distinct model; One provider; shared blind spots possible; Fresh blinded review context not documented for every model; Benchmark match incomplete; Adverse reports or unresolved objections require editorial disposition
OpenAI · Codex (based on GPT-6; exact served backend/version and reasoning effort unknown) · unknown · review context: unknown4.00
Recommendation: Major Revision · Material objections: 3
The exact9-page PDF is a coherent, apparently correct finite technical note: exact orthogonal unrestricted ambient/coexact restoring ranks, a fixed-positive-angle order-n Fourier upper with n-independent row coefficient, a small-angle means-only counterexample, canonical ambient count lower and uncut-cell coexact falsifier. No definite fatal mathematical error was found. Importance4.00/10 is this PDF’s involved AI judgment, not an inherited250B score or physical progress. Exact priority and stand-alone significance beyond a classical-mechanism finite benchmark remain unresolved; recommendation major_revision, distinct from a working-paper deposit or acceptance.
- Correctness confidence
- Read all9actual PDF text pages and individually inspected all9original pagePNGs. Adversarial checks of real curl-adjoint preimages/Nyquist, graph spectrum, exact actual ranks, cutoff/angle normalization, both counterexamples and ambient domain boundary found no definite fatal mathematical error. Analytical AI reading only, not formal verification.
- Rigor
- Explicit original source-weighted labels, period2, actual coexact image, matched injection lower/constructive upper ranks, transformed-norm return and finite cutoff cases. The canonical ambient proof is not wrongly transferred to U; D4 upper is a cited proved source247 input, checked in full. No proof assistant or independent mathematical execution.
- Novelty
- Specific exact formulas and row-class separation are plausible specialist novelty candidates, but historical priority is unestablished. Selected primary Spillane2011/2014 definitions/statements confirm spectral coarse enrichment is established. No exact-equivalence or exhaustive priority audit was performed; self-contained finite calculations alone do not establish originality.
- Significance
- A coherent grouped finite benchmark/technical note with useful exact ranks and bounded falsifiers. One prescribed parity toy profile, nonlocal/mixed unrestricted rows, absent application/solver/global condition-number result and the still-open canonical coexact logarithmic question limit demonstrated stand-alone or wider research importance.
- Reproducibility
- Exact355193bytePDF/source/hash binding, all9pages inspected, full source/delta/engineering reports and cited247 source read. Matching clones and27Linux child exits are attributed retained engineering evidence, not my own rerun. Current sourceZIP/TeX hashes checked; not all115source/110input files newly read or scientifically executed.
Strengths
- Actual source-weighted forward labels, period-two conventions and counting-normalized observations are explicit.
- The special strong-component projection’s coexact image and dimension are proved with genuine real curl-adjoint preimages, including Nyquist and zero n2endpoint.
- Matched low-energy injection and explicit mean upper determine both orthogonal unrestricted minimum ranks.
- Positive-angle cutoff rows are actual original-field functionals, with orthogonality used in Y followed by the singular-value return; no unproved U-invariance is needed.
- Small-angle and uncut-cell witnesses are real coexact fields with the advertised kernel properties; their scopes are properly bounded.
- The D4 canonical/unrestricted contrast compares the same ambient domain and coefficients, while the canonical coexact logarithmic question remains explicitly open.
Weaknesses
- Exact historical priority and predecessor implication for the specialized formulas are unestablished after bounded primary comparisons.
- The general spectral/minimax/enrichment mechanisms are classical; elementary explicit specialization alone does not demonstrate a new broad principle.
- Stand-alone wider relevance remains modest for one parity toy profile with no application, implemented solver or global condition-number result.
- Unrestricted upper rows are nonlocal and may mix channels; constants deteriorate as angle tends to zero and provide no canonical upper.
- Canonical lower is ambient only; the uncut-cell necessary condition and known logarithmic upper do not resolve the actual coexact canonical minimum.
- A compact domain/class/rank/count/angle/penalty table and precise packaged link to cited247 would improve navigation; historical production/review wording needs later sidecars rather than score inheritance.
Potential errors
- No definite fatal mathematical error or counterexample was found within the exact finite stated hypotheses; this is not a formal or exhaustive certificate.
- Extraction can obscure the conjugation in equation10 and sqrt2 in page7; the actual page images confirm the required symbols, so these are not identified PDF errors.
- Rectangular averaging has not been shown to preserve U. The manuscript correctly limits equation12 to ambient restoration; applying that kernel proof on U would be invalid.
- The fixed-angle theorem does not extend exact orthogonal minima or remain uniform as angle tends to zero; current text states these limits.
Strong novelty candidates
- Exact orthogonal ambient rank Dn versus actual coexact rank(D-1)(n-2), with a derived preserving component projection, as a specialist finite formula subject to priority scrutiny.
- Same-profile D4 ambient canonical order n2 versus unrestricted exact4n, using the proved canonical upper from cited247, subject to exact-priority judgment.
- Explicit fixed-positive-angle transverse cutoff and the real small-angle means-only falsifier as a transparent finite benchmark, rather than a claimed new spectral-enrichment principle.
Unresolved material objections
- 255B-M1 | OPEN | material | Originality/priority: Exact historical novelty of the specialized formulas and rectangle obstruction remains unestablished. Classical spectral/coarse-mode principles are not new merely because explicitly specialized; limited primary reading proves neither novelty nor prior implication.
- 255B-M2 | OPEN | material | Stand-alone significance/practical relevance: Coherent as one companion note, but wider importance beyond a finite worked example is not demonstrated: no application, implemented solver, full-operator conditioning or resolution of the canonical coexact log question. The n-independent unrestricted K is not a practical efficiency theorem.
- 255B-M3 | OPEN | material | Scope/editorial interpretation; no fatal finite defect found: Retain orthogonal D3/D4 exact minima versus fixed-angle asymptotic order, ambient versus actual coexact domains, unrestricted versus canonical rows and actual rank versus count. Neither negative example settles arbitrary canonical O(n) partitions. No physical/continuum, formal, independent peer-review or unconditional acceptance claim follows.
arr:assessment:aea24750-e947-40b2-b05d-cda22918393b · prompt ARR-ASSESS-1.0 · response SHA-256 a2652d082b1cd73999bdd2e4a1827c1c33094de1e9f5ebedc6d006be19833d03 · canonical PDF SHA-256 d45a17a4142204628189441b9133aa1027b0d7577a56dc5de6c576d2af6102e7
OpenAI · Codex (based on GPT-6; exact served backend/version and reasoning effort unknown)
Original AI-only score: 4 / 10 · Scientific importance0-10; overall4of10, criteria4/4/2/2/4 of5
major_revision
First exact255B PDF assessment, no rescore; all3material objections OPEN. Originality and wider relevance unestablished. Not acceptance.
Backend identity: unknown. Involved nonblind shared-account authoring/handling context; prior250B score visible, not transferred. Exact served backend/version/effortunknown. Model assessments and review context retained in full.
Report captured · Exact PDF SHA-256 d45a17a4142204628189441b9133aa1027b0d7577a56dc5de6c576d2af6102e7
- assessment255B-first-round/primary-reading-notes.txt · SHA-256
f7910ad51208ed871edfcce1f533bb71bb67254d4c83a63358bfd26c201ab266 - assessment255B-first-round/prompt.txt · SHA-256
e71fa12f66e9a3a27e700efffd87083a4f013ea57fabeeea8d91eec3cb2a825d - assessment255B-first-round/report-hashes.json · SHA-256
5350750cd6ce1d743ead26606b15b5bd910b2c8d32751add7c02a3ff43bd28b0 - assessment255B-first-round/response-native.json · SHA-256
7da6ba3645478e3278723f4ade1de4c72a093ceb33813735ef218bc4a020ff18 - assessment255B-first-round/review-context.json · SHA-256
4bd70b3b6409410f02bcd2d21a5489e0d863b011ad13d1691cd861387b75ea2c - assessment255B-first-round/scientific-review.txt · SHA-256
bba99630ea8a461efd3766531568203c8d4feb2c22e28f434ffdfcbeb10d30c5 - Author response and dated later context · SHA-256
4cd00e0784166225760db8fc5caaa4abd78cf0a68569dc6837fad3dc8395c1f6
A model assessment is not peer review or a correctness certificate. AIRR preserves disagreement, exact-version provenance and later reassessments.
Editorial disclosure
One grouped finite research Working paper with a modest technical-note contribution. First involved exact-PDF AI4.00/10,4stars,criteria4/4/2/2/4, major_revision and all3material objections OPEN. Clean native receipt and actual CC-BY/CC0 permission; no human signoff, acceptance or independent peer review.