{"id":"bec5c2d8-4f9f-4bda-89a0-6c62894d8e94","arxiv_id":"2412.15686","paper_version":1,"verdict":"REJECT","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"high","formal_verification":"none","parameter_count":0,"one_line_summary":"For curves, surfaces with q=pg=0, and hypersurfaces of dimension 2 or 3, the paper gives criteria for when Ulrich bundles are projectively normal, with a likely error in the hypersurface determinant computation.","lead":"Smooth projective varieties built from Ulrich vector bundles are tested for a strong embedding property called projective normality. The paper proves numerical criteria on curves and surfaces and claims that on low-dimensional hypersurfaces Ulrich bundles are rarely projectively normal.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified: the reader's determinant conflict is an arithmetic slip (on a degree-d surface in P3, K_S=(d-4)H, so K_S+3H=(d-1)H), and Theorem 3's determinant hypothesis agrees with Remark 2.4.","rationale":"The reader's verdict of REJECT rests on the claim that Theorem 3(a) uses a determinant formula r/2(d-1)H that contradicts Remark 2.4. That claim is arithmetically false: for a smooth surface of degree d in P3, K_S=(d-4)H, so K_S+3H=(d-1)H, and the determinant assumption in Lemma 7.3 is exactly what Remark 2.4 predicts. I therefore checked the surrounding computation rather than the alleged contradiction. Lemma 7.3's formulas follow from Lemma 6.6, and the resulting 2-normality thresholds are correct by a direct dimension count. Lemma 7.4's threefold formulas are also internally consistent; my independent rank-2 cubic threefold check reproduced both chi(E tensor E)=36 and chi(S2E)=21. Because the only stated reason for rejection is invalid and I could not identify a different load-bearing flaw in the hypersurface theorem, honesty requires a non-finding. I set verdict_should_be to UNCHANGED solely because no concern emerged that would move the verdict under the provided rules; the paper's central claim appears technically sound on the points examined.","tokens_in":880,"tokens_out":858,"duration_ms":284422,"concrete_test":"Verify Lemma 7.4(iv)-(v) on a concrete rank-2 Ulrich bundle on a smooth cubic threefold X in P4: solve c2(E) from chi(E)=6 (this forces c2(E)=5/3H^2), then compute chi(E tensor E) and chi(S2E) by Hirzebruch-Riemann-Roch using the resolution (7.2). The expected values are 36 and 21, respectively; any deviation would indicate a Chern class or HRR error in the threefold part.","verdict_should_be":"UNCHANGED","load_bearing_attack":"I checked the reader's load-bearing objection and found it does not land. For a smooth surface S in P3 of degree d, adjunction gives K_S=(d-4)H, hence K_S+3H=(d-1)H. Remark 2.4 therefore gives c1(E)=r/2(K_S+3H)=r/2(d-1)H, exactly the determinant hypothesis used in Lemma 7.3 and Theorem 3(a). The reader's alternative c1(E)=rd/2H would only follow from K_S=(d-3)H, which is not the correct canonical class. I also rechecked the numerical consequences: substituting the correct K_S into Lemma 6.6 reproduces the formulas in Lemma 7.3, and comparing h0(S2E)=rd/24(d+1)((d+5)r+6) with dim S2H0(E)=rd(rd+1)/2 gives precisely the stated thresholds (d=2, or d=3 with r>=3, or d=4 with r>=6). For the threefold case, Lefschetz gives Pic(X)=Z and c1(E)=r/2(d-1)H; the Chern class and Euler characteristic formulas in Lemma 7.4 are internally consistent, for example on a rank-2 cubic threefold where the resolution (7.2) and Hirzebruch-Riemann-Roch give chi(E tensor E)=36 and chi(S2E)=21, matching (iv) and (v). I did not find a load-bearing error in the central hypersurface argument.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the projective normality of the projective bundle P(E) of an Ulrich vector bundle E, i.e., the normal generation of the tautological line bundle O_{P(E)}(1). The main results are: on curves, projective normality and the (N_p) property hold under degree conditions on the polarization (Theorem 1); on surfaces with q=p_g=0, a degeneracy-locus characterization of non-normal, non-aCM 0-regular bundles is given (Theorem 2); and on hypersurfaces of dimension 2 and 3, it is shown that Ulrich bundles are often not projectively normal under a determinant assumption (Theorem 3). The paper is organized around Castelnuovo-Mumford regularity of tensor operations, Chern class computations, and degeneracy loci, and it contains substantial technical work in the appendices.","tokens_in":32429,"tokens_out":57846,"duration_ms":461744,"significance":"If the main results are correct, the hypersurface statements are noteworthy: they show that very ample Ulrich bundles, which exist in abundance on line-free hypersurfaces, need not be projectively normal, contrary to a naive expectation from the curve case. The surface characterization in Theorem 2 is a useful structural result, and the paper provides a battery of explicit Chern class and Euler characteristic computations for tensor and symmetric powers of Ulrich bundles. The paper is written in a clear, self-contained style, and many proofs reduce to cited results in a transparent way. A particular strength is the careful treatment of regularity of tensor powers under different hypotheses on the polarization.","major_comments":[{"comment":"The formulas in Lemma 7.4 produce non-integral values for some parameter pairs, which is impossible for Chern numbers and Euler characteristics. For example, when d=4 and r=3, formula (i) gives c_3(E)=99/4, formula (iv) gives chi(E⊗E)=315/2, and formula (v) gives chi(S^2E)=315/4. These pairs cannot support an Ulrich bundle, and indeed c_1(E)=r/2(d-1)H is not Cartier when r(d-1) is odd. The lemma and the proof of Theorem 3(b) should explicitly impose the integrality condition (for instance r(d-1) even) or state that the formulas also show non-existence in the remaining cases. As written, the proof of Theorem 3(b) uses a non-integral chi in the inequality h^0≥chi>dim, which is not a valid numerical argument for an existing bundle.","section":"Lemma 7.4 and Theorem 3(b)"},{"comment":"The step from the vanishing (7.3) for E⊗E to the analogous vanishing for S^2E is implicit. Since S^2E is a direct summand of E⊗E over the complex numbers, the vanishings for E⊗E imply those for S^2E; this should be stated explicitly, otherwise the computation of chi(X,S^2E)=h^0-h^1 is not fully justified.","section":"Proof of Lemma 7.4(v)"},{"comment":"The range r>(d+4)/3 in Theorem 3(b) includes parameter values for which the determinant r/2(d-1)H is not an integral divisor and hence no Ulrich bundle can exist. The theorem is vacuously true in those cases, but the presentation should either restrict to admissible ranks or add a remark that the Chern class formulas already rule out those values. This is important because a reader may otherwise believe that a rank-3 Ulrich bundle on a quartic threefold exists but is merely non-normal, whereas in fact such a bundle cannot exist.","section":"Theorem 3(b) statement"}],"minor_comments":[{"comment":"When writing c_1(E)=r/2(K_X+(n+1)B) under Pic(X)≅Z, it would be helpful to note explicitly that the right-hand side must be an integral Cartier divisor, i.e., r(K_X+(n+1)B) must be divisible by 2 in Pic(X).","section":"Remark 2.4"},{"comment":"The notation S^{m-3}E for m=1 appears in the statement; it should be clarified that S^kE=0 for k<0, or the statement should be split according to dimension.","section":"Proposition 4.18"},{"comment":"The sentence 'the assumption on the Picard group forces r≥2' is terse; it would be clearer to add a one-line justification that a rank-1 Ulrich bundle would have h^0(E)=d, whereas O_S(kH) for k=(d-1)/2 has a different h^0 on a general surface.","section":"Proof of Lemma 7.3"},{"comment":"Many symbols in the text appear corrupted (e.g., '/shortrightarrow' for arrows), and there are occasional missing spaces. These should be corrected in the final version.","section":"General typography"}],"recommendation":"major_revision","confidential_remarks":"The core mathematical content appears sound after the determinant issue raised by the stress-test is resolved: adjunction gives K_S=(d-4)H on a surface in P^3, so the determinant used in Lemma 7.3 is consistent with Remark 2.4. The main revision needed is to address the integrality and existence issues in Lemma 7.4 and Theorem 3(b), which currently allow impossible parameter values to enter the proofs. I would also suggest the editor verify the status of the companion preprint [But], since it is cited for a key ampleness statement on curves."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: the rejection based on an internal determinant contradiction is wrong. For a smooth surface S in P3 of degree d, K_S=(d-4)H, so K_S+3H=(d-1)H. Remark 2.4 therefore gives c1(E)=r/2(d-1)H, which is precisely the hypothesis in Lemma 7.3 and Theorem 3(a). The reader's alternative c1(E)=rd/2H would only follow from K_S=(d-3)H. I rechecked the numerical consequences: substituting the correct K_S into Lemma 6.6 reproduces Lemma 7.3, and the thresholds (d=2, or d=3 with r>=3, or d=4 with r>=6) come out exactly as stated. For the threefold, the Chern class and Euler characteristic formulas in Lemma 7.4 are internally consistent; on a rank-2 cubic threefold the resolution (7.2) gives chi(E⊗E)=36 and chi(S2E)=21, matching (iv) and (v). So the central hypersurface argument holds up.\n\nWhat is genuinely new: Theorem 1 gives concrete degree bounds for projective normality and (Np) of Ulrich bundles on curves, extending Butler and Green–Lazarsfeld to the vector bundle setting. Theorem 2 is a genuinely new geometric characterization of projective normality failure on regular surfaces with pg=0, in terms of degeneracy loci of sections of Λ2M_E*. Theorem 3 is a real structural result: on low-dimensional hypersurfaces, Ulrich bundles are almost never projectively normal, contrary to the naive expectation from very ampleness. The paper is carefully written and the proofs mostly reduce cleanly to known theorems.\n\nSoft spots, in proportion: the paper is long and dense, and the surface criterion, while elegant, is not easy to apply because it requires controlling degeneracy loci and a divisor in |K_S+(h−r−1)E|. Some sharpness claims are conditional on the existence of Ulrich bundles with prescribed determinant and rank, which remains conjectural in general; the author notes this, but readers should not take the numerical bounds as unconditional existence statements. Theorem 3(b) gives a sufficient rank bound, not a classification; Remark 7.5 shows low-rank cases behave differently, so the full picture on hypersurfaces is still open. These are limits, not defects.\n\nWho this is for: algebraic geometers working on Ulrich bundles, syzygies, and projective normality. It extends known line-bundle results to vector bundles in low dimensions in a way that is both useful and credible. I would bring it to a reading group and would cite it if I were working in this area.\n\nRecommendation: send it to peer review. The main claims are correct as far as I can tell, and the paper deserves referee time. I would suggest minor revisions mainly for organization and for making the conditional nature of some existence-dependent statements more explicit.","headline":"The reader's main objection is an arithmetic slip: on a degree-d surface in P3, adjunction gives c1(E)=r/2(d-1)H, exactly as used in Lemma 7.3. The paper is sound on its central claims and deserves a serious referee.","tokens_in":32984,"tokens_out":2138,"would_cite":true,"duration_ms":20852,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14J60","14N05","14H60","13D02"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper establishes that Ulrich bundles on smooth hypersurfaces of dimension two or three are almost never projectively normal, with explicit degree and rank thresholds.","keywords":["Ulrich bundles","projective normality","projective bundles","hypersurfaces","Castelnuovo-Mumford regularity","syzygy bundles","degeneracy loci","vector bundles on curves"],"falsifier":"Find a smooth quintic surface $S\\subset\\mathbb{P}^3$ carrying an Ulrich bundle $E$ of rank $r\\ge 2$ with $\\det(E)=O_S(2rH)$ such that the multiplication map $H^0(S,E)\\otimes H^0(S,E)\\to H^0(S,E\\otimes E)$ is surjective. Lemma 7.3 and Theorem 3(a) predict this map cannot be surjective, so one such example would refute the central claim.","tokens_in":1914,"feed_emoji":"📐","tokens_out":2625,"duration_ms":89902,"temperature":0.7,"pith_summary":"The paper asks when an Ulrich vector bundle, embedded through the complete linear system of its tautological line bundle, gives a projectively normal variety. It shows that on curves and on surfaces with $q=p_g=0$ the behavior is governed by explicit degree and degeneracy conditions, while on smooth hypersurfaces of dimension 2 and 3 Ulrich bundles are almost never projectively normal. On such a hypersurface of degree $d$ in $\\mathbb{P}^{n+1}$, the multiplication map on sections fails to be surjective in all but small-rank cases, so the tautological embedding is not normally generated. This matters because Ulrich bundles are globally generated and, on line-free hypersurfaces, very ample; the paper shows that their projective embeddings nevertheless tend to have non-normal coordinate rings.","feed_headline":"Ulrich bundles rarely give normal embeddings on hypersurfaces","feed_subtitle":"On 2- and 3-dimensional hypersurfaces, Ulrich bundles fail projective normality except in narrow low-rank cases.","key_machinery":"The central object is the projective bundle $\\mathbb{P}(E)$ over $X$ with tautological line bundle $\\mathcal{O}_{\\mathbb{P}(E)}(1)$; projective normality of $E$ means that the maps $S^kH^0(X,E)\\to H^0(X,S^kE)$ are surjective for all $k$. The paper's main tools are the multiplication map $\\mu_E:H^0(E)\\otimes H^0(E)\\to H^0(E\\otimes E)$ and the syzygy bundle $M_E=\\ker(H^0(E)\\otimes \\mathcal{O}_X\\to E)$. On curves and surfaces, Castelnuovo-Mumford regularity of tensor powers is controlled through $M=\\max\\{1,\\mathrm{reg}_B(\\mathcal{O}_X)\\}$, keeping symmetric powers $0$-regular, so cohomology vanishings translate into normality. On hypersurfaces, the Ulrich resolution $0\\to E(-d)\\to \\mathcal{O}_X(-1)^{\\oplus rd}\\to \\mathcal{O}_X^{\\oplus rd}\\to E\\to 0$, combined with Chern class formulas for $c_2$ on surfaces and $c_3$ on threefolds, computes $h^0(E\\otimes E)$ and $h^0(S^2E)$. The dimension inequality $\\dim S^2H^0(E) < h^0(S^2E)$ then rules out $2$-normality. For surfaces, degeneracy loci of sections of $\\Lambda^2M_E^*$ encode failure of normality through a zero-dimensional scheme $Z$ lying on a divisor $D$.","core_discovery":"The paper's central claim is Theorem 3: for a smooth hypersurface $X\\subset\\mathbb{P}^{n+1}$ of dimension $n=2$ or $3$, Ulrich bundles are rarely projectively normal. For $n=2$, assuming $\\det(E)=O_X(\\tfrac{r}{2}(d-1)H)$, the section multiplication map cannot be surjective, and $E$ cannot be projectively normal, when $d\\ge 5$, or $d=4$ and $r\\le 5$, or $d=3$ and $r\\le 2$. For $n=3$ and $d\\ge 4$, the multiplication map is never surjective, and $E$ cannot be projectively normal when $r>\\tfrac{d+4}{3}$. The paper also proves that on curves of genus $g$, a $B$-Ulrich bundle is projectively normal when $\\deg B>g+1$ and satisfies higher syzygy properties for larger degree, and that on surfaces with $q=p_g=0$, failure of projective normality is equivalent to the existence of a zero-dimensional degeneracy locus $Z$ lying on a divisor $D$ in a prescribed linear system. The overall message is that the naive expectation linking ample or very ample Ulrich bundles to projective normality fails in higher dimensions.","pith_inferences":["The same dimension-count method could be pushed to higher-dimensional hypersurfaces once the relevant Chern classes and the regularity of symmetric powers are controlled, likely yielding an analogous rank threshold for non-normality.","The surface equivalence in Theorem 2 suggests a constructive route to non-normal Ulrich bundles: choose a zero-dimensional scheme $Z$ and a divisor $D$ satisfying the stated conditions, then build a bundle $E$ whose syzygy bundle has $Z$ as degeneracy locus.","If Theorem 3's qualitative conclusion persists under corrections to the determinant hypothesis, then the general Ulrich bundle in moduli on a hypersurface should be expected to be non-projectively normal, making the syzygies of general Ulrich embeddings genuinely complicated."],"forward_implications":["On a smooth curve of genus $g$, every $B$-Ulrich bundle is projectively normal as soon as $\\deg B>g+1$, and satisfies the higher syzygy property $(N_p)$ for sufficiently large degree.","On a smooth surface with $q=p_g=0$, failure of projective normality of an ample $0$-regular bundle is equivalent to a concrete geometric condition: a zero-dimensional degeneracy locus $Z$ contained in a divisor from $|K_S+(h-r-1)\\det(E)|$.","On smooth surfaces in $\\mathbb{P}^3$ of degree $d\\ge 5$, no Ulrich bundle satisfying the stated determinant condition can be projectively normal; on threefold hypersurfaces with $d\\ge 4$, the section multiplication map is never surjective.","The heuristic that very ample Ulrich bundles on line-free hypersurfaces should be projectively normal is false: such bundles are very ample yet almost never normally generated.","Projective normality of Ulrich bundles is an open property in flat families, so projectively normal Ulrich bundles form open subsets of the relevant moduli spaces."],"supporting_citations":[{"why":"Provides the locally free resolution of Ulrich bundles on hypersurfaces that underlies the Chern class and section-dimension computations.","marker":"[Tr1, §2]"},{"why":"Gives the standard linear resolution of an Ulrich bundle, used for the syzygy and tensor-power analysis.","marker":"[Be2, Proposition 2.1 & Theorem 2.3]"},{"why":"Supplies the second Chern class formula for Ulrich bundles on surfaces used in Lemma 6.6 and Lemma 7.3.","marker":"[C, Proposition 2.1]"},{"why":"Gives the third Chern class of Ulrich bundles on threefold hypersurfaces used in Lemma 7.4.","marker":"[BMPMT, Proposition 3.7]"},{"why":"Establishes the Castelnuovo-Mumford regularity of tensor products on surfaces, keeping symmetric powers 0-regular.","marker":"[S, Lemma 1.4 & Proposition 1.5]"},{"why":"Provides the cohomological criterion converting exact sequences into k-normality statements.","marker":"[L1, Example B.1.3]"},{"why":"Shows Ulrich bundles on line-free hypersurfaces are very ample, framing why the non-normality result is unexpected.","marker":"[LS, Theorem 1]"},{"why":"Gives strong 2-normality for vector bundles on curves, the engine behind Theorem 1(a).","marker":"[Bu, Theorem 2.1]"}],"fun_headline_variants":["Projective normality rare for Ulrich bundles on low-dimensional hypersurfaces","Embeddings of Ulrich bundles rarely projectively normal on hypersurfaces","On hypersurfaces, Ulrich bundles fail projective normality except in low rank","Low-dimensional hypersurfaces: Ulrich bundles rarely projectively normal"],"cache_read_input_tokens":35072,"weakest_assumption_plain":"The hypersurface results assume the standard Chern class formula $c_1(E)=\\frac{r}{2}(d-1)H$ for Ulrich bundles and that the computed Euler characteristics equal the actual dimensions of section spaces because higher cohomology vanishes; if either fails, the numerical obstructions do not apply.","fun_headline_variants_meta":{"raw":{"variants":["Projective normality rare for Ulrich bundles on low-dimensional hypersurfaces","Embeddings of Ulrich bundles rarely projectively normal on hypersurfaces","On hypersurfaces, Ulrich bundles fail projective normality except in low rank","Low-dimensional hypersurfaces: Ulrich bundles rarely projectively normal"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001792,"raw_usage":{"total_tokens":7013,"prompt_tokens":847,"completion_tokens":6166,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":463,"completion_tokens_details":{"reasoning_tokens":6091}},"tokens_in":463,"tokens_out":6166,"duration_ms":37069,"temperature":1.0,"reasoning_tokens":6091,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T11:13:48.730118+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Find a smooth quintic surface $S\\subset\\mathbb{P}^3$ carrying an Ulrich bundle $E$ of rank $r\\ge 2$ with $\\det(E)=O_S(2rH)$ such that the multiplication map $H^0(S,E)\\otimes H^0(S,E)\\to H^0(S,E\\otimes E)$ is surjective. Lemma 7.3 and Theorem 3(a) predict this map cannot be surjective, so one such example would refute the central claim.","supporting_citations":[],"review_version":1}