{"id":"9596efc3-1157-4767-a889-a7abe9090ea2","arxiv_id":"1908.03113","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For several large classes of functions in the infinite polydisk Hardy space, a function is cyclic if and only if it is zero-free, yielding complete solutions to the Periodic Dilation Completeness Problem in those cases.","lead":"This paper proves new necessary-and-sufficient criteria for when functions in the Hardy space over the infinite-dimensional polydisk are cyclic, which translate into completeness criteria for dilation systems in L2(0,1). The results generalize several classical theorems on the Periodic Dilation Completeness Problem and connect to the Riemann hypothesis through Noor's criterion.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lemma 3.5 is false as stated; its proof applies inequality (3.3) to a slice G(szζ) that need not lie in H2(D), and Theorem 3.1's sufficiency proof relies on this lemma.","rationale":"The reader's conditional verdict is reasonable, but the single most load-bearing problem is internal rather than the external [DG] coextension theorem: Lemma 3.5, used in the proof of the central Theorem 3.1, is false as stated, and the proof of the lemma applies inequality (3.3) to a slice that is not known to be in H2(D). This is a concrete, checkable defect that does not depend on any outside preprint. It does not by itself refute Theorem 3.1, because in the actual application F is zero-free and a direct repair seems available, but the manuscript as written contains a false lemma and a gap in a central proof. The reader's other concerns, including the overstatement about the Riemann hypothesis and the reliance on [DG], remain secondary. Since the correct response is still conditional acceptance with mandatory revision, I keep the reader's verdict unchanged.","tokens_in":33556,"tokens_out":27240,"duration_ms":293732,"concrete_test":"Verify the stated counterexample: take F(ζ)=ζ1−1/2, G=1/F, r=3/4, and compute FG_r. Since the denominator vanishes at ζ1=2/3∈D, FG_r is not in H2∞, disproving Lemma 3.5 as stated. Then either repair Lemma 3.5 by adding a zero-free hypothesis and rewrite the proof of Theorem 3.1 without invoking the false lemma, or supply a correct proof of the specific estimate used in the zero-free case.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Lemma 3.5 asserts: if F∈A_{R,∞} (R>1), F(0)≠0, G is holomorphic on D∞2, and FG∈H2∞, then FG_r∈H2∞ for every 0<r<1. This is false. Take R=2, F(ζ)=ζ1−1/2, G=1/F. Then F(0)=−1/2≠0, G is holomorphic on D∞2 because |ζ1|<1, and FG=1∈H2∞. But for r=3/4, FG_r=(ζ1−1/2)/((3/4)ζ1−1/2), whose denominator vanishes at ζ1=2/3∈D; hence FG_r is not holomorphic on D∞2, let alone in H2∞. The proof fails at the displayed estimate below (3.3): inequality (3.3) is applied with h(z)=G(szζ), but G(szζ) need not belong to H2(D). Only the product F(szζ)G(szζ) is known to be an H2 slice. Since the proof of Theorem 3.1 uses Lemma 3.5 with G=1/F, the proof of the sufficiency direction is incomplete as written. The zero-free case is probably repairable: for fixed r, 1/F_r is bounded on D∞, so F/F_r∈H2∞ and the required weak convergence follows from boundedness and pointwise convergence. But the lemma as stated and its proof must be corrected.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies cyclic vectors in the Hardy space H^2_∞ over the infinite-dimensional polydisk, viewed through the Bohr transform as the function-theoretic avatar of the Periodic Dilation Completeness Problem. The main results are: a zero-free criterion for functions in the infinite polydisk algebra A_{R,∞} (Theorem 3.1); a characterization of cyclicity for infinite products of functions in independent variable sets (Theorem 4.2); a composition theorem for independent inner functions (Theorem 5.1); a criterion showing that if FG is cyclic with F in H^p_∞ (p>2) and G in H^q_∞ (q>0), then F is cyclic (Theorem 6.1); and a geometric condition involving the image of F avoiding a curve with bounded argument (Theorem 7.1). The final section applies the machinery to the Kozlov completeness problem for dilation systems of χ_{(0,θ)}.","tokens_in":33821,"tokens_out":38003,"duration_ms":411016,"significance":"If the central theorems are correct, they constitute a substantial advance: they unify and generalize earlier results of Nikolski, Hartman, Kozlov, and Hedenmalm–Lindqvist–Seip, and they give the first complete cyclicity criteria for several large natural classes of functions on the infinite polydisk. The proofs in Sections 3–7 are detailed, largely self-contained, and use no fitted parameters; several arguments, especially the product theorem and the H^p interpolation theorem, are elegant and convincing. However, the application in Section 8 contains a false Fourier identity that invalidates the proof of the Kozlov claims as written, and the composition theorem depends on a coextension theorem quoted from an arXiv preprint. These issues are local and repairable, but they must be fixed before the paper is publishable.","major_comments":[{"comment":"The paper states that for ψ(x)={x}−1/2 one has ψ(x)=∑_{n=1}∞ −(1/(nπ)) sin(nπx), and then uses this to derive the identities φ_1=2ψ(2x)−4ψ(x), φ_{1/2}=ψ(4x)−2ψ(x)−ψ(2x), etc. These statements are not correct for the stated ψ. The Fourier sine series −∑ (1/(nπ)) sin(nπx) equals (x−1)/2 on (0,1), not {x}−1/2. For example, with ψ={x}−1/2, the function 2ψ(2x)−4ψ(x) equals +1 on (0,1/2) and −1 on (1/2,1), whereas χ_{(0,1)} is identically 1. Consequently the subsequent formulas F_θ=P_θ K_p and the proof that D_θ is complete for θ=1,1/2,2/3 and incomplete for θ=1/3 are not valid as written. The authors should either change the definition of ψ to (x−1)/2 throughout Section 8, or recompute all Fourier coefficients and the resulting identities from the actual definition of ψ.","section":"Section 8, Eq. (8.1) and the preceding display"},{"comment":"The proof of Theorem 5.1 depends in an essential way on the theorem from [DG] asserting that a doubly commuting sequence of C·0-contractions has a coextension to a doubly commuting sequence of pure isometries with V_n^*|_H=T_n^*. This theorem is quoted from an arXiv preprint (arXiv:1907.05815) but is not proved in the present manuscript. Since Proposition 5.4, Corollary 5.6, and hence the necessity direction of Theorem 5.1 all rest on it, the authors should either include a proof of the coextension theorem in the paper or replace the citation with a published peer-reviewed reference. As it stands, a central link in the main chain of Section 5 cannot be checked from the manuscript alone.","section":"Section 5, Proposition 5.4 and Corollary 5.6"}],"minor_comments":[{"comment":"The sentence 'the argument in the previous paragraph shows FG_{ts}∈H^r_∞ for any 0≤s≤1' is not literally correct: interpolation gives FG_{ts}∈H^{u_s}_∞ with 1/u_s=(1−s)/p+s/r, hence u_s≥r, not necessarily u_s=r. What is actually needed later is only 1/u_s≤1/r, which does hold, so the proof can be repaired by a short clarification.","section":"Section 6, proof of Theorem 6.1(1)"},{"comment":"The displayed identity −(√2/(2π))BUψ=∑_{n=1}∞(−1)^n(1/n)ζ^{α(n)} does not match the normalization ψ_n=√2 sin(nπx): for ψ(x)=x the Fourier coefficients give BUψ=(2√2/π)∑(−1)^{n+1}(1/n)ζ^{α(n)}, so the scalar should be −π/(2√2), not −√2/(2π). The cyclicity conclusion is unaffected by a nonzero scalar, but the formula should be corrected.","section":"Example 4.8"},{"comment":"The problem is named after Kozlov, and the section title and several occurrences read 'Kolzov'. Please correct the spelling.","section":"Throughout Section 8"},{"comment":"The proof says 'Put R=2^ε'; this is fine, but the estimate |ζ^{α(n)}|≤R^{α_1+...+α_m}≤n^ε after setting R=2^ε depends on the inequality p_j^{ε α_j}≥2^{ε α_j}, which should be stated explicitly for clarity.","section":"Corollary 3.7"},{"comment":"The proof of Proposition 2.3 cites [Ru, Theorem 3.4.3] for the inclusion H^p(D^n)⊆H^p_∞; it would help the reader if the precise statement and the identification of functions depending on finitely many variables were spelled out at that point.","section":"Lemma 2.2 reference"}],"recommendation":"major_revision","confidential_remarks":"The core Sections 3–7 appear to contain sound and significant mathematics, and the issues I found are concentrated in Section 8 and in the external dependence on [DG]. I am not recommending rejection because the main cyclic-vector theorems are independent of the Section 8 computation and are likely correct. I would ask the editor to ensure that, in the revision, the authors either prove or provide a published reference for the coextension theorem from [DG], and that all Fourier identities in Section 8 are recomputed from the definition of ψ actually used."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper is a substantial, serious piece of work on cyclicity in H^2_∞ and hence on the PDCP. The stress-test note that came with it does not hold up. The proposed counterexample to Lemma 3.5 takes G=1/(ζ1−1/2), which is not holomorphic on D∞2 because of the pole at ζ1=1/2. The lemma's hypotheses are not met. The step in the proof that worried the stress-test, applying (3.3) to G(szζ), is legitimate: for fixed ζ in the compact set overline D^n and s<1, the slice G(szζ) is bounded on the disk and hence in H^2(D). So Lemma 3.5 stands.\n\nThe new results are real. Theorem 3.1 gives a zero-free criterion for the infinite polydisk algebra A_{R,∞}, extending Nikolski's finite-variable theorem, and Corollary 3.7 is a nice concrete consequence. Theorem 4.2 for products in independent variables and Theorem 4.4 for partition-multiplicative coefficients are natural generalizations of Hartman–Kozlov, with transparent proofs. Theorem 5.1 on composition with inner functions is the most involved; its necessity argument uses a projection identity that depends on the coextension theorem from the authors' preprint [DG]. That dependence is load-bearing, and [DG] should be published or its proof put in an appendix. The use is not circular: it is a dilation-theoretic statement, independent of cyclicity.\n\nThe genuine soft spots are rhetorical. The abstract's claim to have solved the PDCP 'in almost all interesting cases' overstates the scope, since the paper treats specific classes and even suggests a universal criterion is out of reach. More seriously, the introduction says Section 6 implies a weak version of the Riemann hypothesis is true. Section 6 does not show that. The authors carefully note there that density of (I−S)span{W_n h_m} is known, but density of span{W_n h_m} would be needed for RH, and that remains open; the phrase 'weak version of RH' is never defined. This needs to be rewritten.\n\nWho is this for? Specialists in Hardy spaces over infinite polydisks, dilation theory, and Dirichlet series. It should be sent to a serious referee. My own verdict is conditional: the math looks sound, the theorems are substantial, but the authors should tone down the claims and make the reliance on [DG] explicit before publication.","headline":"A substantial advance on cyclicity in H^2_∞ with three genuine new theorems; the stress-test's counterexample to Lemma 3.5 is invalid, and the paper's only real weaknesses are overclaims about RH and scope.","tokens_in":34419,"tokens_out":5762,"would_cite":true,"duration_ms":54485,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["42C30","47A16","46E50","46E22","42B30"],"pacs":[],"model":"deepseek-v4-flash","headline":"Zero-free functions on the infinite polydisk are cyclic, and cyclicity survives products and inner compositions.","keywords":["Periodic Dilation Completeness Problem","cyclic vectors","Hardy space over the infinite polydisk","Bohr transform","Dirichlet series","multiplication operators","Kozlov completeness problem","inner functions"],"falsifier":"Compute the Fourier-Bohr coefficients of $G_\\theta = F_\\theta \\cdot K_p^{-1}$ for a rational value such as $\\theta=\\frac15$; Proposition 8.1 predicts that $G_\\theta$ depends on infinitely many variables for every $\\theta$ outside $\\{1,\\frac12,\\frac13,\\frac23\\}$. If a finite computation shows all but finitely many coefficients vanish, the paper's Kozlov completeness classification would be wrong.","tokens_in":33297,"feed_emoji":"🔁","tokens_out":9938,"duration_ms":106981,"temperature":0.7,"pith_summary":"The paper's target is a classical completeness problem from the 1940s: for which functions $\\psi\\in L^2(0,1)$ is the integer dilation system $\\{\\psi(kx):k\\ge1\\}$ dense in $L^2(0,1)$? Through the Bohr transform this becomes the question of which functions are cyclic vectors in the Hardy space $\\mathbf{H}^2_\\infty$ over the infinite-dimensional polydisk, and the paper gives necessary-and-sufficient answers for most of the function classes that have been studied. Its main theorems say that zero-free functions in the infinite polydisk algebra $A_{R,\\infty}$ with $R>1$ are cyclic; that an infinite product of functions in independent variables is cyclic exactly when every factor is; and that composing with mutually independent inner functions preserves cyclicity. If correct, these criteria solve the periodic dilation completeness problem on those classes and cover nearly all earlier results.","feed_headline":"Zero-free means cyclic in the infinite polydisk","feed_subtitle":"Necessary-and-sufficient criteria settle a 1940s dilation completeness problem for most function classes.","key_machinery":"The carrier of the argument is the Hardy space $\\mathbf{H}^2_\\infty$, the Hilbert space of power series in infinitely many variables with square-summable coefficients, together with the Bohr transform that turns a Dirichlet series or an odd periodic function's Fourier series into such a power series. In that setting, a function is cyclic when the joint invariant subspace generated by the coordinate multiplications $M_{\\zeta_n}$ is the whole space, which is equivalent to completeness of the corresponding dilation system. The proofs run on three mechanisms: the infinite polydisk algebra $A_{R,\\infty}$ with a local-boundedness estimate for multiplication by dilated functions; a product criterion based on pairwise independent variables and orthogonal projections onto finite variable sets; and a composition operator $C_\\Phi(F)=F(\\varphi_1,\\varphi_2,\\ldots)$ whose boundedness and range are controlled by a coextension theorem for doubly commuting contractions, giving the projection identity (5.3).","core_discovery":"On the paper's own terms, the central discovery is that cyclicity in $\\mathbf{H}^2_\\infty$ admits complete characterizations for several large classes. Theorem 3.1 states that for $R>1$ and $F\\in A_{R,\\infty}$, $F$ is cyclic if and only if $F$ has no zero in $\\mathbb{D}^2_\\infty$. Theorem 4.2 states that if $\\{F_n\\}$ are functions with mutually independent variables and the infinite product converges pointwise to a nonzero $F\\in \\mathbf{H}^2_\\infty$, then $F$ is cyclic if and only if each $F_n$ is cyclic. Theorem 5.1 states that if $\\{\\eta_n\\}$ are nonconstant inner functions with mutually independent variables and $\\sum_n |\\eta_n(0)|<\\infty$, then $F(\\eta_1,\\eta_2,\\ldots)$ is cyclic if and only if $F$ is cyclic. These results are used to characterize cyclicity for functions whose coefficients satisfy multiplicative conditions relative to a partition of the primes, for composite functions, and for $\\mathbf{H}^p_\\infty$ functions with $p>2$, with applications to the Kozlov completeness problem.","pith_inferences":["The zero-set criterion for $A_{R,\\infty}$ is proved with a quantitative constant depending on the number of zeros on a smaller radius; a natural extension would be to functions with merely absolutely summable coefficients, where the same Hurwitz argument might survive without the $n^\\varepsilon$ growth condition.","The product theorem suggests a practical algorithm for cyclicity testing: factor a Dirichlet series by prime groups and check outer-ness in one variable per group, since numerical zero location in one variable is much easier than invariant-subspace approximation in infinitely many variables.","The paper stops short of proving its conjecture that $D_\\theta$ is complete only for $\\theta\\in\\{1,\\frac12,\\frac23\\}$; computing $G_\\theta$ for a rational like $\\theta=\\frac15$ is a concrete test of that boundary.","The Riemann hypothesis appears only as a downstream remark; if the density claims for the $\\phi_m(z^n)$ systems could be strengthened to the full span of the associated weighted composition semigroup, the methods would connect directly to the Nyman-type completeness criterion for the Riemann hypothesis."],"forward_implications":["Functions in $A_{R,\\infty}$ with $R>1$ and no zeros in $\\mathbb{D}^2_\\infty$ are cyclic, so for Dirichlet series whose coefficients decay faster than $n^{-\\varepsilon}$ for some $\\varepsilon>0$, completeness of the dilation system is decided entirely by the zero set.","An infinite product of cyclic factors with independent variables is cyclic, and only cyclic factors can produce a cyclic product; this gives a prime-factorization reduction for functions with multiplicative coefficients.","Cyclicity is invariant under composition with mutually independent inner functions, so cyclicity of $1+\\sum_n a_n\\eta_n$ is equivalent to $\\sum_n |a_n|\\le 1$.","For the Kozlov completeness problem, the systems $D_\\theta$ with $\\theta=1,\\frac12,\\frac23$ are complete, $D_{1/3}$ is not, and the remaining cases reduce to a finite-variable check that the paper conjectures selects exactly those three values."],"supporting_citations":[{"why":"Supplies the coextension theorem for doubly commuting C.0-contractions that controls the composition operator and yields the projection identity (5.3).","marker":"[DG]"},{"why":"Establishes the unitary correspondence between L^2(0,1) dilation systems and the Dirichlet-series Hardy space, the starting point for the cyclicity reformulation.","marker":"[HLS]"},{"why":"Provides the zero necessary condition, the zero-free sufficiency for functions holomorphic near the finite closed polydisk, and reproducing kernel estimates used throughout.","marker":"[Ni1]"},{"why":"Gives the original cyclicity results for totally multiplicative and multiplicative coefficients that Section 4 generalizes.","marker":"[Har]"},{"why":"Independent multiplicative-coefficient characterization, the baseline for the partition-multiplicative theorems.","marker":"[Koz2]"},{"why":"Builds the Hardy-space theory over the infinite polydisk, including evaluation estimates and density of A_infty, used in Sections 2 and 5.","marker":"[CG]"},{"why":"Polynomial zero-free cyclicity and the classical inequality used in Lemma 3.3's estimate.","marker":"[NGN]"},{"why":"Provides the weighted composition semigroup whose completeness consequences Proposition 6.6 strengthens.","marker":"[No]"},{"why":"Formulates the dilation completeness problem independently and points to its arithmetical significance.","marker":"[Win]"}],"fun_headline_variants":["Zero-free means cyclic for most classes","Cyclicity in infinite polydisk: zero-free is key","Beurling-Wintner problem cracked for most classes","Infinite polydisk Hardy space: cyclic vectors classified"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The necessity proof in the composition section rests on an earlier theorem that every doubly commuting family of strongly stable contractions can be embedded in a doubly commuting family of pure isometries on a larger space; if that embedding theorem fails, the projection identity (5.3) breaks and the cyclicity equivalence for composite functions collapses.","fun_headline_variants_meta":{"raw":{"variants":["Zero-free means cyclic for most classes","Cyclicity in infinite polydisk: zero-free is key","Beurling-Wintner problem cracked for most classes","Infinite polydisk Hardy space: cyclic vectors classified"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001541,"raw_usage":{"total_tokens":6189,"prompt_tokens":991,"completion_tokens":5198,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":607,"completion_tokens_details":{"reasoning_tokens":5133}},"tokens_in":607,"tokens_out":5198,"duration_ms":40152,"temperature":1.0,"reasoning_tokens":5133,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:24:56.928038+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the Fourier-Bohr coefficients of $G_\\theta = F_\\theta \\cdot K_p^{-1}$ for a rational value such as $\\theta=\\frac15$; Proposition 8.1 predicts that $G_\\theta$ depends on infinitely many variables for every $\\theta$ outside $\\{1,\\frac12,\\frac13,\\frac23\\}$. If a finite computation shows all but finitely many coefficients vanish, the paper's Kozlov completeness classification would be wrong.","supporting_citations":[],"review_version":1}