{"id":"53e1a5ca-747f-4e8f-a1ba-cb6b0cabfab0","arxiv_id":"2501.14938","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"Sidon sets yield weighted 2-designs of size d^2 + O(d^{1.525}) in every dimension, the first general bound matching the lower bound up to lower-order terms.","lead":"This paper uses dense Sidon sets to build small weighted projective 2-designs in every dimension, achieving the first general upper bound of size d^2 + O(d^{1.525}), close to the d^2 lower bound. It marks quantitative progress on Zauner's conjecture about symmetric quantum measurements.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified: Theorem 7's algebra checks out, and the Baker–Harman–Pintz prime-gap input is a valid published theorem; a weaker gap would only weaken the exponent.","rationale":"The reader marked reliance on Baker–Harman–Pintz as the weakest assumption. I scrutinized exactly that point. It is a published theorem, not an assumption introduced by this paper, and it is used correctly: the smallest prime at least d is at most d+O(d^{0.525}), so p(d)^2 = d^2+O(d^{1.525}); adding d preserves the error term. I also re-checked the entrywise computation of X and E; the Sidon condition is exactly what forces the support of X to be the union of diagonal and off-diagonal matched pairs, giving the claimed weighted decomposition. The numerical and table claims are consistent with the construction. No internal inconsistency or missing proof was found. The reader and I both identify BHP as the only external dependency, but I do not treat it as a load-bearing concern because it is a valid theorem. Since the central claim is well-supported, the ACCEPT verdict stands.","tokens_in":8995,"tokens_out":13673,"duration_ms":128059,"concrete_test":"Write a short exact-arithmetic script (e.g., SymPy) that forms the matrices X, E, and P from Theorem 7 for the Sidon set S = {(x,x^2): x ∈ F_5} in G = F_5^2 and verifies P = (|S|^2/(2|G|))X + (1/2)E; if the identity fails, the central construction has a hidden error.","verdict_should_be":"UNCHANGED","load_bearing_attack":"I could not identify a load-bearing flaw. The core computation in Theorem 7 is correct: for a Sidon set S, the entrywise sums give X = (|G|/|S|^2)(A+B), E = B, and P = (1/2)A+B; hence P = (|S|^2/(2|G|))X + (1/2)E with nonnegative weights. Corollary 8 then gives n(d) ≤ m(d)+d. For sufficiently large d, the Erdős–Turán Sidon set in F_p^2 with p=p(d) gives m(d) ≤ p(d)^2, and Baker–Harman–Pintz indeed supplies p(d) ≤ d+O(d^{0.525}); finite exceptional dimensions are absorbed by the O constant. The only external input is a standard published theorem, so it is not an unproved assumption of this paper; weakening it would degrade the O(d^{1.525}) exponent but would not threaten the construction or the o(d^4) claim. Section 4's explicit limitation that this method cannot reach d^2 is consistent with the theorem as stated.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper constructs weighted projective 2-designs in CP^{d-1} of size d^2 + O(d^{1.525}) for every d, giving the first general upper bound on the minimal design size n(d) that is o(d^4). The construction applies the Bodmann-Haas idea to Sidon sets: Theorem 7 proves directly that for any finite abelian group G and Sidon set S in G, the |G|+|S| unit vectors x_alpha (alpha in hat G) and e_r (r in S) form a weighted 2-design for CP^{|S|-1}, with nonnegative weights |S|^2/(2|G|) and 1/2; the proof is an explicit entrywise computation of the projection onto the symmetric tensors. Corollary 8 then gives n(d) <= m(d) + d, where m(d) is the minimal size of a group containing a Sidon set of size d. Combining the Erdos-Turan Sidon set (x,x^2) in F_p^2 with the Baker-Harman-Pintz prime gap bound p(d) <= d + O(d^{0.525}) yields Theorem 4. The paper also tabulates the resulting bounds for d <= 150, compares them with all previous bounds, and candidly explains why the method cannot reach Zauner's conjecture.","tokens_in":124,"tokens_out":63627,"duration_ms":592061,"significance":"If the results stand, Theorem 4 is a genuine quantitative advance: it is the first dimension-uniform bound of the form d^2 + o(d^2), improving on the previous universal bound (binom(d+1,2))^2 = Theta(d^4) and approaching the SIC lower bound n(d) >= d^2. The proof is a clean, fully checkable direct computation with no fitted parameters, and the external inputs (classical Sidon families and the published Baker-Harman-Pintz theorem) are independent of the paper's claims. The explicit nonnegative weights and the honest statement of limitations in Section 4 are additional strengths. Spot-checks of Table 1 are consistent with the formulas in Proposition 3. One caveat: Proposition 9(c) misstates the Bose-Chowla construction, and since Table 1 and parts of Section 3 use that family, this component needs correction; Theorem 4 itself does not depend on it.","major_comments":[{"comment":"Proposition 9(c) is false as stated. For the relative trace Tr_{F_{q^2}/F_q}(x) = x + x^q, the set S = {x in (F_{q^2})^x : Tr(x) = 0} has exactly q-1 elements, not q, because the trace-zero set is a 1-dimensional F_q-subspace that includes 0. It is also not Sidon in the multiplicative group (F_{q^2})^x: for odd q >= 3, any nonzero trace-zero omega satisfies (-omega)^2 = omega^2 with -omega != omega, so the distinct unordered pairs {-omega,-omega} and {omega,omega} have equal product; for even q >= 4 the set is F_q^x, where 1*g^2 = g*g for g != 1. The intended Bose-Chowla Sidon set of size q is the affine hyperplane {x in (F_{q^2})^x : Tr(x) = c} for a fixed nonzero c in F_q, for instance c = 1; the standard verification is that if a,b,c,d in S satisfy ab = cd and t := a/c = d/b, then applying the trace to a = tc and d = tb gives c(t - t^q) = 1 - t^q and b(t - t^q) = 1 - t^q, forcing either t = 1 or b = c, and hence {a,b} = {c,d} as multisets. This correction is load-bearing for Table 1, whose rows using Bose(q) (for instance d = 2, 7, 11, 13, 16, 19, 23, 25, 29, 31, 37, 41, 43, 47, 49, 53, 59, 61, 64, 71, 73, 79, 81, 83, 89, 97, 101, 103, 107, 109, 113, 121, 125, 127, 131, 137, 139, 149) presume a Sidon set of size q in a group of order q^2 - 1, and for the discussion's assertion that m(d) is never achieved by the Erdos-Turan family. Theorem 4 is unaffected because its proof uses only Proposition 9(a) and the prime-gap bound, but Proposition 9(c) must be corrected and the dependent sentences and table entries re-verified accordingly.","section":"Section 3, Proposition 9(c)"}],"minor_comments":[{"comment":"The sentence 'Corollary 8 is the sharpest known upper bound on n(d) for all but finitely many d' is imprecise, because for the infinite families of dimensions with d-1 a prime power or d a prime power the bound only ties Proposition 3(c) and (d), respectively; please rephrase as 'at least as sharp as, and for all but finitely many dimensions strictly sharper than, the previously known bounds'.","section":"Section 3, after Corollary 8"},{"comment":"The claim 'For every d <= 150, exactly one of three things happens' depends on the completeness of the list of known SIC dimensions taken from the website [8] plus four added references; since the paper does not reproduce the dimension set, the reader cannot check the rows marked (a). Please include the full list of dimensions with known SICs or state explicitly how the list was assembled.","section":"Table 1"},{"comment":"For readability, the proof should add one sentence explaining that the Baker-Harman-Pintz bound on consecutive prime gaps implies p(d) <= d + O(d^{0.525}) for p(d) the least prime >= d, by applying the gap bound to the largest prime strictly below d.","section":"Proof of Theorem 4"}],"recommendation":"major_revision","confidential_remarks":"The reader's report and the stress-test both concluded that no flaw exists in the paper. My reading agrees that Theorem 7's computation and the proof of Theorem 4 are correct, but it also turned up a genuine error that neither caught: Proposition 9(c) defines the Bose-Chowla set with the wrong trace value, giving a set of size q-1 that is not Sidon in (F_{q^2})^x. The fix is standard and local (replace trace 0 by a fixed nonzero trace value), and Theorem 4 does not use it, so I recommend major revision rather than rejection. You may also want the authors to make the SIC dimension list for Table 1 self-contained rather than relying on a website."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper does something genuinely new: it shows that any Sidon set in any finite abelian group gives a weighted projective 2-design via the Bodmann–Haas construction, and then it leverages the densest known Sidon-set families to get n(d) ≤ d^2 + O(d^{1.525}) for every dimension d. That is the first general upper bound that is o(d^4), let alone within lower-order terms of the d^2 lower bound. The connection to Hughes Sidon sets in particular is new and, as the table shows, it drives most of the concrete improvements below dimension 150.\n\nWhat the paper does well: Theorem 7 is a direct computation and it is correct. The averaging over characters gives X = (|G|/|S|^2)(A+B), the standard-basis part gives E = B, and the weights are nonnegative. Corollary 8 then makes the whole strategy transparent: n(d) ≤ m(d)+d, where m(d) is the smallest group with a Sidon set of size d. The use of Erdős–Turán sets and Baker–Harman–Pintz is clean and the O(d^{1.525}) follows immediately. I also appreciate that the authors are upfront in Section 4 that this approach cannot reach d^2, since the construction always gives at least d^2+1 points. That is a real limitation and they state it plainly.\n\nThe soft spots are mostly minor. The main external input, Baker–Harman–Pintz, is a deep analytic number theory theorem, but it is a published theorem and not something the paper needs to reprove. If a weaker prime-gap bound were used, the exponent would degrade but the construction would still be o(d^4), so the core contribution does not stand or fall on that theorem. The table is large and the choices of which Hughes subset to use are computational, but they are reproducible from the stated parameters. One small thing: the text says the weighted design has size |G|+|S|, and the lower bound from the pigeonhole principle gives |G| ≥ d^2−d+1, so indeed |G|+|S| ≥ d^2+1; the paper's claim that the method cannot prove Zauner's conjecture is accurate.\n\nWho is this for? Anyone working on weighted 2-designs, entanglement-breaking rank, or quantitative approaches to Zauner's conjecture. It is a solid, honest paper with a correct central proof and a meaningful new bound. It deserves a serious referee and should be published, probably with only minor revisions. I would cite it if I were writing in this area.\n\nMy recommendation: send it to peer review; it will pass.","headline":"First general nearly-tight weighted 2-design bound via Sidon sets; clean proof, only real dependency is a deep prime-gap theorem.","tokens_in":9725,"tokens_out":902,"would_cite":true,"duration_ms":10154,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["05B30","11B75","42C15"],"pacs":[],"model":"deepseek-v4-flash","headline":"A single Sidon-set construction gives nearly minimal weighted 2-designs in every dimension.","keywords":["weighted 2-designs","complex projective spaces","Sidon sets","Zauner's conjecture","SIC-POVMs","entanglement breaking rank","prime gaps","finite abelian groups"],"falsifier":"Numerically test the construction on a dense Sidon set of size $d$ not covered by Table 1, for example the Erdős–Turán set in $(\\mathbb{F}_p)^2$ with $p=10^4+7$: form the $|G|+|S|$ vectors, compute the weighted sum of rank-one projections, and check equality with the symmetric projection; one mismatch for any Sidon set would disprove the paper's main reduction $n(d)\\le m(d)+d$, while agreement supports it.","tokens_in":8811,"feed_emoji":"🎯","tokens_out":13926,"duration_ms":117892,"temperature":0.7,"pith_summary":"Zauner's conjecture asserts that every complex projective space $\\mathbb{CP}^{d-1}$ carries $d^2$ equally spaced points, a SIC, which is equivalent to the smallest weighted $2$-design having size $n(d)=d^2$. The paper proves $n(d)\\le d^2+O(d^{1.525})$ for every dimension $d$, the first general upper bound that is $o(d^4)$ and therefore within a lower-order term of the $d^2$ lower bound. The proof converts any Sidon set of size $d$ in a finite abelian group $G$ into an explicit weighted $2$-design of size $|G|+d$, then uses dense Sidon sets in groups of order near $d^2$ together with the cited prime-gap estimate. A dimension-by-dimension table through $d=150$ shows the construction improving on all previous bounds except where a SIC is already known.","feed_headline":"Near-minimal quantum designs in every dimension","feed_subtitle":"New Sidon-set construction lands within a lower-order term of the d² minimum, closing in on Zauner's conjecture.","key_machinery":"The machine that carries the argument is a map from any subset $S$ of a finite abelian group $G$ to a sequence of $|G|+|S|$ unit vectors in $\\mathbb{C}^{|S|}$: for each character $\\alpha$ of $G$, take $x_\\alpha(s)=\\alpha(s)/\\sqrt{|S|}$; for each $r\\in S$, take the standard basis vector $e_r$. The proof shows that when $S$ is a Sidon set (all pairwise sums $a+b$ with $a,b\\in S$ distinct), the sum of rank-one projections from the characters equals $(|G|/|S|^2)(A+B)$ and the basis vectors supply $B$, so the symmetric projection $P$ on $(\\mathbb{C}^{|S|})^{\\otimes 2}$ is the nonnegative combination $P=\\frac{|S|^2}{2|G|}X+\\frac{1}{2}E$; that is exactly a weighted $2$-design. Dense Sidon sets with $|G|=d^2+O(d^{1.525})$ come from the families collected in Proposition 9, and the least-prime bound converts the group order into the theorem's exponent.","core_discovery":"The central claim is Theorem 4: $n(d)\\le d^2+O(d^{1.525})$, where $n(d)$ is the smallest size of a weighted $2$-design for $\\mathbb{CP}^{d-1}$ (equivalently, the entanglement-breaking rank of a certain quantum channel). The route is Corollary 8, $n(d)\\le m(d)+d$, with $m(d)$ the smallest order of a finite abelian group containing a Sidon set of size $d$. Applying the group-character construction of Section 2 to any Sidon set gives a weighted $2$-design of size $|G|+|S|$; choosing $S$ of size $d$ from a dense Sidon set in a group of order essentially $d^2$, and bounding the least prime $p(d)\\ge d$ by $d+O(d^{0.525})$ via the cited prime-gap theorem, yields the theorem. This is the first known universal upper bound on $n(d)$ that is $o(d^4)$.","pith_inferences":["Extending the paper's construction, the same Sidon-set pipeline plausibly yields weighted $k$-designs for larger $k$ by replacing the symmetric projection with the projection onto $k$-th order symmetric tensors, giving small tomographically complete measurement sets in every dimension.","Because the table shows Hughes-type Sidon subsets providing the strict improvements, it is likely that multiplicative Sidon sets in groups of the form $(\\mathbb{F}_q^\\times)^2$ are the most efficient raw material known; examining larger $q$ would test whether this advantage persists asymptotically.","The prime-gap dependence couples a problem in quantum information (entanglement breaking rank) to the short-interval distribution of primes; a lower bound on $n(d)$ of the form $d^2+\\omega(d)$ would therefore constrain the density of Sidon sets in finite abelian groups, a connection the paper leaves implicit."],"forward_implications":["For every dimension $d$, a weighted $2$-design of size within $O(d^{1.525})$ of the lower bound $d^2$ can now be written down explicitly, not merely shown to exist.","In all dimensions $d\\le150$ displayed in Table 1, the new bound is the best known unless an exact SIC is already available, and the strict improvements come from taking subsets of the Hughes-type Sidon sets.","The construction cannot prove Zauner's conjecture: any group carrying a Sidon set of size $d$ has order at least $d^2-d+1$, so these designs always have size at least $d^2+1$.","Better prime-gap results would immediately sharpen the exponent, while reaching $d^2+o(d\\log^2 d)$ would require a fundamentally different approach under the usual heuristic for prime gaps."],"supporting_citations":[{"why":"Supplies the prime-gap estimate $p(d)\\le d+O(d^{0.525})$ that turns the Sidon-set group-order bound into the theorem's exponent.","marker":"[4]"},{"why":"Introduces the group-character construction of weighted projective 2-designs that the paper extends to arbitrary Sidon sets.","marker":"[6]"},{"why":"Provides the catalogue of dense Sidon-set families (with parameters) on which Proposition 9 and the bound on $m(d)$ depend.","marker":"[7]"},{"why":"Gives the earlier character-based construction of mutually unbiased bases that the group-character design construction generalizes.","marker":"[9]"},{"why":"Identifies $n(d)$ with the smallest size of a weighted 2-design and supplies the previous universal upper bound $\\binom{d+1}{2}^2$.","marker":"[12]"},{"why":"Establishes $n(d)\\ge d^2$ with equality exactly when a SIC exists, connecting the upper bound to Zauner's conjecture.","marker":"[14]"},{"why":"States the original conjecture that $d^2$ equidistant points exist in every $\\mathbb{CP}^{d-1}$, the target this paper approaches.","marker":"[18]"}],"fun_headline_variants":["Sidon sets build near-tight 2-designs in all dimensions","Near-minimal weighted 2-designs from dense Sidon sets","Sidon construction yields 2-designs within O(d^1.525) of optimum","Sidon sets narrow the gap to Zauner's conjecture"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The bound's stated exponent rests on the cited prime-gap theorem that the smallest prime $p(d)\\ge d$ satisfies $p(d)\\le d+O(d^{0.525})$ for every $d$; the paper does not prove this analytic number theory input.","fun_headline_variants_meta":{"raw":{"variants":["Sidon sets build near-tight 2-designs in all dimensions","Near-minimal weighted 2-designs from dense Sidon sets","Sidon construction yields 2-designs within O(d^1.525) of optimum","Sidon sets narrow the gap to Zauner's conjecture"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001124,"raw_usage":{"total_tokens":4584,"prompt_tokens":760,"completion_tokens":3824,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":376,"completion_tokens_details":{"reasoning_tokens":3745}},"tokens_in":376,"tokens_out":3824,"duration_ms":24734,"temperature":1.0,"reasoning_tokens":3745,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T14:45:31.991400+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Numerically test the construction on a dense Sidon set of size $d$ not covered by Table 1, for example the Erdős–Turán set in $(\\mathbb{F}_p)^2$ with $p=10^4+7$: form the $|G|+|S|$ vectors, compute the weighted sum of rank-one projections, and check equality with the symmetric projection; one mismatch for any Sidon set would disprove the paper's main reduction $n(d)\\le m(d)+d$, while agreement supports it.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the prime-gap estimate $p(d)\\le d+O(d^{0.525})$ that turns the Sidon-set group-order bound into the theorem's exponent."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces the group-character construction of weighted projective 2-designs that the paper extends to arbitrary Sidon sets."},{"cited_title":"Eberhard, F","cited_arxiv_id":null,"evidence_quote":"Provides the catalogue of dense Sidon-set families (with parameters) on which Proposition 9 and the bound on $m(d)$ depend."},{"cited_title":"Godsil, A","cited_arxiv_id":null,"evidence_quote":"Gives the earlier character-based construction of mutually unbiased bases that the group-character design construction generalizes."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Identifies $n(d)$ with the smallest size of a weighted 2-design and supplies the previous universal upper bound $\\binom{d+1}{2}^2$."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes $n(d)\\ge d^2$ with equality exactly when a SIC exists, connecting the upper bound to Zauner's conjecture."},{"cited_title":"Zauner, Quantum designs, PhD thesis, U","cited_arxiv_id":null,"evidence_quote":"States the original conjecture that $d^2$ equidistant points exist in every $\\mathbb{CP}^{d-1}$, the target this paper approaches."}],"review_version":1}