{"id":"a023dfa5-1b06-4bfc-a45d-43e2a4fe32d9","arxiv_id":"2509.00556","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"Two subsets of F_2^n are affinely equivalent exactly when their even-zero-sum Venn diagrams admit a cardinality-preserving linear isomorphism.","lead":"This paper proves a new criterion for when two sets of binary vectors can be moved onto each other by an affine transformation: it is enough to compare the sizes of regions in a certain Venn diagram. The criterion is then used to count and classify Sidon sets (quad caps) in binary affine spaces.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 5.7 appears sound; the load-bearing gap is §7's reliance on template completeness from [2], which determines the cap-classification conclusions.","rationale":"The reader's strongest claim, Theorem 5.7, is supported by a self-contained proof that combines the E(S)-subspace characterization (Theorem 2.6) with the Venn-region isomorphism theorem (Theorem 4.5). I checked the key steps: the change-of-coordinates matrix in Theorem 3.16, the construction of g in Theorem 4.5, and the use of the sumset map in Section 5. I found no missing hypothesis or invalid inference in the main equivalence result. The genuinely load-bearing risk is in the application: Section 7 relies on the completeness of templates taken from the authors' companion paper [2]. The text only says that the tables 'cover all possibilities' and does not reproduce a proof of exhaustiveness. If a template were missing, the Venn invariant would still be computed correctly, but the final classification of 10-, 11-, and 12-caps could be incomplete. This matches the reader's weakest assumption. Since the reader already assigned a conditional verdict for this reason, the verdict should remain unchanged. Minor typographical errors and the lack of citations to matroid theory are not correctness issues.","tokens_in":20460,"tokens_out":20341,"duration_ms":245463,"concrete_test":"Independently regenerate the enumeration underlying [2, Sections 5–7]: implement the extended cap basis type search for AG(7,2) or verify the exhaustiveness proof in [2] line by line, checking that for k=10, 11, 12 the complete template list is exactly Tables 4, 5, and 7. If an additional 11-cap template appears, §7.2's conclusion that all 11-caps are equivalent would be false; if the lists match, the application is safe.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The affine-equivalence criterion in Theorem 5.7 is internally coherent; I could not find a counterexample or a missing hypothesis. The load-bearing weakness is downstream: §7 imports from [2, Sections 5–7] the assertion that the extended cap basis type templates in Tables 4, 5, and 7 exhaust all 10-, 11-, and 12-caps of dimension 7. No proof of exhaustiveness is given here. Consequently the conclusions that all 11-caps of dimension 7 are affinely equivalent, and that the 10-caps split into exactly two classes, are only as sound as that external enumeration. If, for example, an 11-cap template with a different even-zero-sum subspace structure existed, Theorem 5.7 would still apply correctly, but the classification would be incomplete. This is an external premise, not an internal inconsistency, and it does not threaten Theorem 5.7 itself.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a Venn-diagram invariant for subsets S of F2^n relative to a linear subspace V of the power set P(S). For a basis of V it partitions S into Venn regions and introduces a vector-space structure on the collection of regions. It proves that this partition and the vector-space structure are independent of the basis of V (Theorem 3.16, Corollary 3.17), and that a bijection S -> T is an affine equivalence exactly when it induces an isomorphism of the even-zero-sum subspaces E(S), E(T) (Theorems 2.6 and 5.6). The main criterion (Theorem 5.7) states that S and T are affinely equivalent iff there is a cardinality-preserving linear bijection between Venn(S,E(S)) and Venn(T,E(T)). The paper applies this to caps (Sidon sets): Section 6 gives a complete classification for caps with size-dimension difference 3, and Section 7 uses templates from [2] to classify 10-, 11-, and 12-caps in dimension 7.","tokens_in":20748,"tokens_out":10881,"duration_ms":125202,"significance":"If the proofs are repaired, the main theorem is a valuable and elegant reduction: affine equivalence of arbitrary subsets of F2^n becomes a finite, parameter-free linear-algebra problem on the even-zero-sum subspace. The paper is explicit and the constructions are concrete, with worked examples and tables. The Section 6 characterization of caps with size-dimension difference 3 is complete and constructive, and the Venn-diagram method demonstrably recovers the known 7-dimensional cap classifications. However, the Section 7 application relies on an external enumeration from the companion paper [2], and the proof of Theorem 3.16 contains a false nonemptiness assertion; both issues need attention before the paper is accepted.","major_comments":[{"comment":"The application to 7-dimensional caps depends on the assertion that the templates reproduced from [2, Sections 5–7] 'cover all possibilities' for 10-, 11-, and 12-caps. No proof of this exhaustiveness is given here. Consequently the conclusions that all 11-caps and all 12-caps of dimension 7 are affinely equivalent, and that the 10-caps split into exactly two classes, are conditional on the completeness of the external enumeration. This does not affect Theorem 5.7, but it is load-bearing for the stated Sidon-set application. Please either supply a proof of exhaustiveness or explicitly state the reliance on [2] with precise pointers to the relevant theorems.","section":"Section 7, Tables 4, 5, 7"},{"comment":"The proof asserts 'every basis vector e_i is nonzero and has v_X(e_i) nonempty'. This is false in general: for S={1,2,3} and V=span({1},{1,2}) with basis X1={1}, X2={1,2}, one has v_X(10)=X1∩X2^c=∅. Thus Lemma 3.13 cannot be applied to these e_i. Since Corollary 3.17 and Theorem 5.7 rely on the coordinate-change formula, the proof needs repair. The statement may be salvageable by choosing a basis of F2^r consisting of vectors whose Venn regions are nonempty (such a basis exists because the distinct incidence columns span F2^r), but the manuscript must supply a correct argument.","section":"Theorem 3.16"},{"comment":"In Lemma 6.2 proof, 'each of a + b ≤ 6' should be '≥ 6'; the same typo appears in Theorem 6.3 proof. Also, 'a, b, and c must have the same cardinality' should be 'same parity'. These are presentation typos in a proof, but they obscure an otherwise correct argument and should be fixed.","section":"Lemma 6.2 / Theorem 6.3"}],"minor_comments":[{"comment":"In the list of Venn regions, 'v_X(001) = v_Y(111) = {a4}' should likely read 'v_X(111) = v_Y(111) = {a4}', since v_X(001) was already listed as {w3}.","section":"Example 3.18"},{"comment":"The sentence 'Figure 5 shows the Venn cardinality diagrams for the three templates in Table 4' should refer to Table 5, not Table 4.","section":"Section 7.2"},{"comment":"The matrices A and B in Sections 7.2 and 7.3 are asserted to permute cardinalities 'appropriately' without a calculation. A short verification, or a reference to an appendix or supplementary file, would make the application easier to check.","section":"Throughout Section 7"}],"recommendation":"major_revision","confidential_remarks":"The manuscript leans heavily on the companion paper [2], which is submitted but not yet published. Since the Section 7 classification is only as strong as the external template enumeration, please ask the authors to either prove the exhaustiveness or explicitly label the result as conditional. The false nonemptiness claim in Theorem 3.16 is a proof gap that should have been caught; it would be prudent to ask for a careful re-verification of all applications of Lemma 3.13."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The core idea is good: affine equivalence of subsets of F2^n can be read off from the linear structure of Venn region cardinalities, and Theorem 5.7 is likely the right way to frame this. Section 6's classification of k-caps with dimension k-3 is clean, concrete, and easy to reproduce; that is a genuine new contribution worth crediting.\n\nBut there is a real bug in the middle. Theorem 3.16 claims that under a change of basis X -> Y of V, the Venn coordinates transform by (M^T)^{-1}. That is false. Take S={1,2,3}, V=span({1},{1,2}), bases X=({1},{1,2}) and Y=({1},{2}). The change-of-basis matrix M is [[1,0],[1,1]], but the coordinate change on Venn regions is multiplication by M, not by (M^T)^{-1}=[[1,1],[0,1]]. The proof breaks because it assumes every basis vector has a nonempty exclusive region, which need not hold. Corollary 3.14 also overstates: a nonempty X in V need not contain exactly 2^{r-1} nonempty regions.\n\nGood news: these errors do not sink Theorem 5.7. Theorem 4.5's converse is proved without the faulty parts, and the basis-independence of addition on Venn regions actually follows directly because v_Y^{-1}∘v_X is always linear. So the main criterion appears sound. Still, the paper as written contains an incorrect theorem and needs revision.\n\nThe second soft spot is Section 7. The classification of 10-, 11-, and 12-caps in dimension 7 assumes the templates from the companion paper [2] are exhaustive. That completeness is not proved or checked here; if a template is missing, the equivalence-class conclusions would be wrong. This is an external premise, not an error in the Venn method, but it should be flagged prominently.\n\nAlso, Theorem 2.6 — affine equivalence iff there is an even-zero-sum-preserving bijection — is a standard matroid fact. The authors should credit prior work on vector matroids over F2 rather than present it as new.\n\nWho is this for? Anyone working on affine equivalence in F2^n, Sidon sets, or caps. The method is transparent, and Section 6 gives a solid hook. I would send it to a serious referee, with a specific request to check Theorem 3.16 and the completeness assumption in Section 7. Fix those, and it becomes a useful paper.","headline":"Main affine-equivalence criterion is sound and useful, but there is a false change-of-basis theorem and Section 7 leans on an unproven external enumeration.","tokens_in":21183,"tokens_out":15073,"would_cite":true,"duration_ms":156921,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["05B25","11B30","51E15"],"pacs":[],"model":"deepseek-v4-flash","headline":"Two subsets of F_2^n are affinely equivalent exactly when their even zero-sum spaces yield cardinality-preserving linearly isomorphic Venn region structures.","keywords":["affine equivalence","Venn diagrams","even zero-sum sets","Sidon sets","caps","finite affine geometry","binary vector spaces","classification"],"falsifier":"An exhaustive computer enumeration over all subsets of F_2^n for n up to 6 (or over all 7-dimensional caps) checking whether every pair of sets is affinely equivalent if and only if a cardinality-preserving linear bijection exists between their Venn region spaces; a single pair with a matching cardinality-preserving linear bijection but no affine equivalence, or vice versa, would disprove Theorem 5.7, and a 7-dimensional cap matching no template from [2] would falsify the application's completeness assumption.","tokens_in":20439,"feed_emoji":"🔢","tokens_out":5164,"duration_ms":60662,"temperature":0.7,"pith_summary":"The paper proves that affine equivalence of subsets of F_2^n can be detected from the Venn diagram of a naturally defined linear space: the collection E(S) of even-size subsets of S that sum to zero. The main theorem states that S and T are affinely equivalent if and only if there is a cardinality-preserving linear bijection between their Venn region spaces. This turns a search for an affine automorphism into a finite linear-algebra comparison. The paper then uses this criterion to classify Sidon sets (caps): it fully classifies caps whose size exceeds dimension by 3, and it verifies the affine equivalence classes of 10-, 11-, and 12-caps in dimension 7.","feed_headline":"Venn diagrams decide affine equivalence of binary sets","feed_subtitle":"Sidon-set caps become equivalent exactly when their Venn region counts match under a linear map.","key_machinery":"The even zero-sum space E(S): the F_2-linear subspace of P(S) (with symmetric difference as addition) containing all even subsets of S that sum to zero in F_2^n. Since affine combinations in characteristic 2 are exactly odd sums, E(S) captures affine dependence, and affine maps are precisely bijections preserving it. The Venn diagram of E(S) is made into an F_2-vector space whose regions are indexed by coordinate vectors; a change of basis in E(S) transforms the region coordinates by the transpose-inverse matrix, so the vector-space structure of the regions is basis-independent. The proof then equates (V,W)-isomorphisms of sets with cardinality-preserving linear bijections of the correspondi","core_discovery":"The central claim is Theorem 5.7: for subsets S,T of F_2^n, there is an affine automorphism taking S to T if and only if there is a cardinality-preserving linear bijection from the Venn diagram of E(S) to the Venn diagram of E(T), where E(S) is the subspace of the power set P(S) consisting of even-cardinality subsets whose elements sum to zero. The discovery is that affine structure is completely encoded by which even subsets sum to zero, and that this algebraic data can be reorganized into a Venn diagram that is independent of the chosen basis. The proof shows that any bijection preserving E(S) extends to an affine map, and that such bijections are exactly the cardinality-preserving linear","pith_inferences":["A natural next step is to test the criterion computationally for n between 8 and 10, where maximum Sidon set sizes are still open, by comparing the Venn cardinality linear structures as an equivalence invariant against known constructions.","The method may be recast as classifying subsets by the isomorphism type of a weighted linear code—the even zero-sum space together with region weights—which could connect the classification to coding-theoretic invariants such as hulls or weight enumerators.","The central mechanism depends essentially on characteristic 2, since affine combinations reduce to parity; an open question is whether an analogous 'weight-restricted zero-sum subspace' invariant characterizes affine equivalence over F_q for odd q.","The Section 7 conclusions inherit an external completeness assumption about the template list; an independent exhaustive enumeration of 7-dimensional caps up to affine equivalence would either confirm or refute the 'all 11-caps equivalent' and 'all 12-caps equivalent' statements."],"forward_implications":["Affine equivalence of any pair of subsets of F_2^n can be decided by comparing the linear structure of Venn region cardinalities, without constructing an explicit affine map.","When the Venn region cardinality multisets differ, the two sets are certainly not affinely equivalent, giving a cheap negative test (Corollary 4.4).","Caps with size minu dimension equal to 3 are classified by two data: the multiset of Venn region cardinalities and the number of isolated points, leading to an explicit parameterization and machine count of equivalence classes up to k=27.","For the 7-dimensional caps carried over from the companion template list, the method reproduces the known classes: 10-caps split into two equivalence classes, while all 11-caps and all 12-caps in the templates are equivalent.","The main theorem is fully general, applying to all subsets—not only caps—so the Venn criterion is a candidate tool for classifying Sidon sets in higher dimensions where complete classification remains open.","The approach yields a canonical linear invariant, E(S), of dimension |S| minus (dim(S)+1), computed directly from any affine basis of S, so no search over automorphisms is needed to state the invariant."],"supporting_citations":[{"why":"Supplies the extended cap basis type templates for 10-, 11-, and 12-caps of dimension 7 that the Section 7 classification relies on for completeness.","marker":"[2]"},{"why":"Establishes the AG(n,2) model, the description of quads as affine planes, and the affine-geometry background used throughout the paper.","marker":"[4]"},{"why":"Provides the definition and context for generalized caps (Sidon sets) in AG(n,q) that the paper's applications target.","marker":"[1]"}],"fun_headline_variants":["Venn diagrams reveal affine equivalence, no basis needed","Sidon sets up to affine equivalence via Venn logic","Venn diagram isomorphism: the affine equivalence test","Even subsets sum to zero: Venn diagram encodes affine map"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The Section 7 classification assumes that the extended cap basis type templates listed in the companion paper [2] include every possible 10-, 11-, and 12-cap of dimension 7; if some cap fits none of those templates, the conclusions that all 11-caps are equivalent (or that the 10-cap classes split as described) could be wrong.","fun_headline_variants_meta":{"raw":{"variants":["Venn diagrams reveal affine equivalence, no basis needed","Sidon sets up to affine equivalence via Venn logic","Venn diagram isomorphism: the affine equivalence test","Even subsets sum to zero: Venn diagram encodes affine map"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00024,"raw_usage":{"total_tokens":1318,"prompt_tokens":674,"completion_tokens":644,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":418,"completion_tokens_details":{"reasoning_tokens":581}},"tokens_in":418,"tokens_out":644,"duration_ms":7980,"temperature":1.0,"reasoning_tokens":581,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T13:28:28.119833+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"An exhaustive computer enumeration over all subsets of F_2^n for n up to 6 (or over all 7-dimensional caps) checking whether every pair of sets is affinely equivalent if and only if a cardinality-preserving linear bijection exists between their Venn region spaces; a single pair with a matching cardinality-preserving linear bijection but no affine equivalence, or vice versa, would disprove Theorem 5.7, and a 7-dimensional cap matching no template from [2] would falsify the application's completeness assumption.","supporting_citations":[{"cited_title":"How Many Cards Should You Lay Out in Quad-128: A Classification of Caps in AG(7,2)","cited_arxiv_id":"2501.11173","evidence_quote":"Supplies the extended cap basis type templates for 10-, 11-, and 12-caps of dimension 7 that the Section 7 classification relies on for completeness."},{"cited_title":"How many cards should you lay out in a game of EvenQuads: a detailed study of caps in AG( n, 2)","cited_arxiv_id":null,"evidence_quote":"Establishes the AG(n,2) model, the description of quads as affine planes, and the affine-geometry background used throughout the paper."}],"review_version":1}