{"id":"21ac7e57-6e0f-40ed-99d4-eae19baea94e","arxiv_id":"2412.04241","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A Habiro ring of a number field is constructed, with K3-graded modules, and perturbative quantum invariants are shown to be elements of these modules.","lead":"This paper introduces a Habiro ring for any number field, together with modules indexed by the third algebraic K-group of the field. It proves that power series from perturbative Chern-Simons theory and from Donaldson-Thomas invariants are elements of these arithmetic objects, connecting quantum topology, enumerative geometry, and number theory.","discovery_kind":"unification","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 5 rests on two unverified inputs: the omitted N>1 case of the q-difference uniqueness proof and an unchecked application of the external integrality theorem [10, Thm.1.6].","rationale":"I read the paper in good faith. The construction of the Habiro ring and its modules is substantial, and the paper contains many independently checkable worked examples. Theorem 5 is the central claim, and its proof has two soft spots: the omitted N>1 case and the reliance on an external integrality theorem. The reader's weakest assumption identified exactly these two points. I do not find grounds to reject the paper: the external theorem is published and the omitted generalization is plausibly routine. However, both points are load-bearing for the stated generality, and the paper itself signals the limitation in Remark 1.9. The numerical checks in Section 4 provide evidence but do not replace a proof for all N and all m prime to Δ. Therefore I agree with the CONDITIONAL verdict and recommend no change. The proposed concrete tests would settle whether the omitted vector case works and whether the constant-term identification is genuinely covered by [10].","tokens_in":65828,"tokens_out":6743,"duration_ms":73015,"concrete_test":"Independently write out the N=2 analogue of equations (213) for A=[[1,1],[1,1]]: define ψ_{μ,ν}(t,q)=F_A(q^{μ_1}t_1,q^{μ_2}t_2,q)F_A(q^{-ν_1}t_1,q^{-ν_2}t_2,q^{-1}) with μ,ν∈Z^2, derive the resulting q-difference equations from (33), and prove by induction on total t-degree that the solution with constant term 1 is unique. If the induction degenerates for t-degrees in a sublattice, the gluing proof of Theorem 5 has a real gap. Separately, for m=5 and p=7 (both prime to Δ for Q(√−3)), compute the constant term of f_{4_1,z,5}(x) from the explicit Gaussian formula (118) and compare it with ε_5(ξ)^{1/5} obtained from the Chern-class recipe of [10]; any discrepancy modulo 5 would show the external theorem is not being applied correctly.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central membership claim fA,z(q) ∈ HR[δ^{-1/2}],ξ|Δ in Theorem 5 depends on two load-bearing inputs that are not fully established in the paper. First, the proof explicitly omits the multi-variable case N>1, declaring it 'completely analogous' in Section 3.3 (and again in the proof of Theorem 8 in Section 2.7). The uniqueness argument for the q-difference system (213) is written only for N=1: the shifts μ, ν are scalars, and the induction on t-degree using the q^k−1 factor has no displayed analogue for the vector shifts in (33). The theorem is then applied to N=2 (the 41 knot, A=[[1,1],[1,1]]) and N=3 (the 52 and (−2,3,7)-pretzel knots), so any hidden obstruction in the vector-valued q-holonomic system would invalidate the gluing condition (24) and hence the module membership. Second, the proof asserts that 'for m prime to ∆' the constant term of fA,z,m(x) is ε_m(ξ)^{1/m} times an integral element, citing [10, Thm.1.6] together with Hutchinson [31]. The hypotheses of [10] are not stated, and the identification of the Gaussian-integral constant term with the K-theoretic unit ε_m(ξ) is not demonstrated in the present text. Remark 1.9 concedes that the stronger statement without the |Δ restriction depends on [10, Thm.1.6] holding for all m, so the validity for the restricted set of m is genuinely load-bearing. If either input fails — the omitted N>1 verification or the applicability of [10] — Corollary 1.10 and the arithmetic integrality computations in Section 4 would not follow.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces, for a number field K with ring of integers O_K and an integer Δ, a 'Habiro ring' H_R consisting of Galois-invariant collections of power series at roots of unity that satisfy a Frobenius-twisted gluing condition, together with rank-one modules H_{R,ξ} indexed by elements ξ of K_3(K). The main results identify the collection of power series coming from admissible Nahm sums with the collection defined by formal Gaussian integration (Theorem 3), establish Dwork-type congruences for these series (Theorem 4), and show that, after specializing t=1 to a non-degenerate solution of the Nahm equations, the resulting perturbative series f_{A,z}(q) belongs to the Habiro module H_{R[δ^{-1/2}],ξ}|_Δ (Theorem 5). The paper also gives an explicit congruence-theoretic description of H_R in Section 5, constructs local sections using infinite Pochhammer symbols and p-adic dilogarithms, and reports extensive numerical verifications for the knots 4_1, 5_2, and (−2,3,7).","tokens_in":66247,"tokens_out":7972,"duration_ms":83788,"significance":"If Theorem 5 is fully established, the paper provides a precise arithmetic home for perturbative complex Chern–Simons invariants and ties Donaldson–Thomas/admissible series to algebraic K-theory and Frobenius-twisted integrality. The framework is novel and the auxiliary results are substantial: the explicit sections Ψ_{[ζ],p}, the Dwork-type lemma (Lemma 3.4), the identification in Theorem 8, and the determinant/congruence analysis in Section 5 are concrete and likely to be reused. The construction is not circular: the modules H_{R,ξ} are defined before the series are introduced, and the main work is verifying the defining conditions. However, two load-bearing inputs are not established in the submitted text: the multi-variable uniqueness argument for the q-difference systems used in Theorem 8 and in the proof of Theorem 5 is omitted, and the applicability of the external theorem [10, Thm. 1.6] to the constant terms of the Gaussian-integral series is asserted without a full statement of its hypotheses or a deduction. The significance is therefore conditional on these points being supplied.","major_comments":[{"comment":"The uniqueness argument for the q-difference system is written only for N=1. In Section 2.7 the proof says 'The proof in the general case is identical using equation (137), and is omitted,' and in Section 3.3 the corresponding statement says 'We again omit the case when N > 1, since it is completely analogous.' This is load-bearing because Theorem 5 is applied in Section 4 to N=2 (the 4_1 knot) and N=3 (the 5_2 and (−2,3,7)-pretzel knots). The scalar recursion in Eq. (213) uses a one-dimensional induction on the power of t and the invertibility of factors such as (q^{k+1−m};q)_m; for N>1 the shifts act componentwise and the corresponding coefficient system must be shown to have a unique formal solution in Q[ζ]( (x) )[[t]] by a multi-index induction. This is plausible but is not automatic, and the reader cannot verify the claimed uniqueness from the displayed text. Please supply the N>1 argument or reduce it explicitly to the scalar case.","section":"Section 2.7, proof of Theorem 8; Section 3.3, proof of Theorem 5, Eq. (213)"},{"comment":"The proof asserts that, for m prime to Δ, the constant term of f_{A,z,m}(x) is ε_m(ξ)^{1/m} times an integral element, citing [10, Thm. 1.6] and Hutchinson [31]. The hypotheses of [10, Thm. 1.6] are not stated, and the identification of the constant term of the formal Gaussian integral with the K-theoretic unit ε_m(ξ) is not demonstrated in this paper. Remark 1.9 explicitly notes that the unrestricted statement would follow only if [10, Thm. 1.6] holds for all m, so the validity for the restricted set of m prime to Δ is genuinely load-bearing. The cited theorem is published and hence not circular, but the paper must state the exact hypotheses and explain why they are satisfied in the present setting, including which primes are excluded and why they are covered by the factor Δ in Eq. (42).","section":"Section 3.3, first paragraph of the proof of Theorem 5; Remark 1.9"},{"comment":"The integer Δ is not specified precisely enough for the statement of Theorem 5. Definition 1.1 only requires Δ to be divisible by the discriminant (and usually by 6), while the proof of Theorem 5 says that Δ 'includes the primes 2 and 3 and finitely many other primes that depend only on the number field K.' Since the notation H_R|_Δ and H_{R,ξ}|_Δ restricts to roots of unity of order prime to Δ, the truth of (42) depends on which additional primes are included. If those primes are not explicitly determined, the theorem's statement is ambiguous and Corollary 1.10 and the computations in Section 4 cannot be checked against a fixed theorem. The authors should either make Δ a precise function of K (or of A and z) or state the theorem with an explicit hypothesis on the primes excluded by Δ.","section":"Section 1.4, Definition 1.1; Section 3.3, proof of Theorem 5"}],"minor_comments":[{"comment":"The sentence 'Added subsection 1.1 explaining what the paper is about and subsection 1.8 explaining the relation to perturbative complex Chern-Simons theory' appears in the manuscript and should be removed before publication; it is a revision note rather than part of the paper.","section":"Front matter, after the abstract"},{"comment":"The statement 'mN mU FGI_m(t)2m ∈ S(m)' appears garbled; presumably it should read something like 'm^N U_m^{FGI}(t)^{2m} ∈ S(m)' or 'm^N m U_m^{FGI}(t)^{2m} ∈ S(m)'. Please correct the notation.","section":"Section 2.5, Lemma 2.12"},{"comment":"There is a typo: 'as q appraoches roots of unity' should read 'as q approaches roots of unity.'","section":"Section 1.6, first sentence"},{"comment":"There is a typo: 'Since the nuymber D(N)' should read 'Since the number D(N).'","section":"Section 5.1, paragraph after Proposition 5.1"},{"comment":"References [23] and [24] appear to be duplicates: both list Garoufalidis and Zagier, 'Asymptotics of Nahm sums at roots of unity', Ramanujan J. 55 (2021), 219–238. If a different article was intended for one of these citations, please correct it.","section":"References [23] and [24]"},{"comment":"The sentence 'If this is true, then 7 f_P belongs to HZ[1/3] and its image under ι to HZ[1/3]' is missing a verb; it should say 'and its image under ι belongs to HZ[1/3]' (or 'to H_{Z[1/3]}' if that is what is meant).","section":"Section 5.4, Example 5.9"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is ambitious and contains many valuable explicit constructions. The main blockers are not mathematical novelty but missing details in the proof of Theorem 5: the omitted N>1 uniqueness argument and the unstated hypotheses of [10, Thm. 1.6]. Both are probably fixable within the scope of the paper, especially since the scalar case is written in detail and the numerical evidence in Section 4 is extensive. I would encourage the editor to request a revision in which these two points are addressed explicitly, rather than rejecting the paper, provided the authors can supply the missing arguments."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know this paper is the real thing. It introduces the Habiro ring of a number field and rank-one modules over it indexed by K3(K), then proves that the series coming from perturbative Chern-Simons theory and from Kontsevich-Soibel'man admissible series land in those modules. The single-variable arguments are careful and detailed, and the worked examples (4_1, 5_2, the (-2,3,7)-pretzel) make the picture concrete. The equality of admissible series and formal Gaussian integration series (Theorem 3) is genuinely new, and Theorem 5 — the membership of f_{A,z}(q) in HR[δ^{-1/2}],ξ|Δ — is the payoff. The construction in Section 3.2 of explicit sections using Pochhammer symbols is elegant, and the congruence description in Section 5 gives a useful alternative handle.\n\nThe soft spots are exactly where the stress-test note points, and they are real but not fatal. Theorem 8 and Theorem 5 both omit the N>1 case, saying the proof is identical or completely analogous. The scalar shifts in equation (213) are not the vector shifts of (33), and since the theorem is applied to N=2 and N=3, a referee should ask to see the vector-valued uniqueness argument spelled out. The second issue is the reliance on [10, Thm.1.6] for the constant term of f_{A,z,m}(x). The hypotheses are not stated and the identification with ε_m(ξ) is asserted. Choosing Δ to avoid bad primes helps, but the paper should state explicitly what [10] requires and confirm those conditions hold. Both gaps are addressable; the single-variable proof gives a clear template, and [10] is published with independent proofs, so this is not a circular problem.\n\nWho should read this? Number theorists, quantum topologists, and anyone working on q-de Rham cohomology or Donaldson-Thomas invariants. It deserves a serious referee. I recommend sending it to review, with the request that the authors fill in the multi-variable details and state the external theorem precisely. The absence of code for the reported computations is a minor issue for a pure math paper, not a blocker.","headline":"A genuinely new arithmetic home for perturbative quantum invariants, with a real but addressable gap in the multi-variable proofs.","tokens_in":66764,"tokens_out":5518,"would_cite":true,"duration_ms":54698,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["11R70","19F27","33D15"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper constructs a Habiro ring for every number field and places perturbative quantum invariants of knots and 3-manifolds into modules over it labelled by the algebraic K-group $K_3(K)$.","keywords":["Habiro ring","number fields","algebraic K-theory","Bloch group","p-adic dilogarithm","admissible series","infinite Pochhammer symbol","perturbative Chern-Simons invariants"],"falsifier":"Take the Nahm data for the $5_2$ knot, whose field is the cubic field of discriminant $-23$, and compute the level-5 series $f_{A,z,5}(x)$ to, say, 200 terms by formal Gaussian integration. Divide by the explicit Pochhammer generator formed from the order-24 root of unity described in the paper's Example 4.3; membership in the module requires every coefficient to be 5-integral, with no negative power of 5 in the denominator. A single coefficient with negative 5-adic valuation would refute the claimed inclusion.","tokens_in":65651,"feed_emoji":"🪢","tokens_out":10297,"duration_ms":101622,"temperature":0.7,"pith_summary":"The paper builds a new arithmetic home for perturbative quantum invariants: a Habiro ring attached to a number field $K$, together with modules over it labelled by the algebraic K-group $K_3(K)$. An element of the ring is a compatible collection of power series around every complex root of unity that glue $p$-adically after a Frobenius twist; the modules are the same collections with constant terms twisted by an $m$-th root of a K-theoretic unit $\\varepsilon_m(\\xi)$. The main theorem states that the perturbative Chern$-$Simons series $f_{A,z}(q)$, defined for a symmetric integer matrix $A$ and a non-degenerate solution $z$ of the Nahm equations with associated Bloch-group class $\\xi$, lies in the Habiro module $H_{R[\\delta^{-1/2}],\\xi}$ restricted to roots of unity of order prime to $\\Delta$. If correct, the asymptotic expansions of quantum invariants of knots and 3-manifolds have a precise arithmetic meaning, and Donaldson$-$Thomas invariants acquire an arithmetic one.","feed_headline":"Knot invariants find a home in K-theory of number fields","feed_subtitle":"Power series at roots of unity p-adically glue via Frobenius, putting asymptotic knot invariants into K3-graded Habiro modules.","key_machinery":"The central object is the Habiro ring $H_R$ of a number field: the set of collections $(f_m(x))_{m\\ge 1}$ with $f_m(x)\\in R[\\zeta_m][\\![x]\\!]$, $R=O_K[1/\\Delta]$, satisfying the Frobenius-twisted gluing condition $f_m(x+\\zeta_{pm}-\\zeta_m)=(\\varphi_p f_{pm})(x)$ after $p$-adic completion. The modules $H_{R,\\xi}$ are the same collections with leading term $\\varepsilon_m(\\xi)^{1/m}$ and with a logarithmic condition that places the $p$-adic dilogarithm $D_p(\\xi)$ in the polar part. The engine of the explicit construction is the infinite Pochhammer symbol $(t;q)_\\infty=\\prod_{n\\ge 0}(1-q^n t)$: its two complementary expansions and the associated Dwork-type difference identity are what turn analytic series into $p$-adically integral and Frobenius-glued ones. The matching of the two sources of series is carried by the $q$-holonomic system whose classical limit is the $t$-deformed Nahm equations.","core_discovery":"On its own terms, the paper claims that the combinatorial data of an integral symmetric matrix $A$ and a non-degenerate solution $z$ of the Nahm equations produces a collection of power series $f_{A,z,m}(x)$ at each $m$-th root of unity that is an element of the rank-one Habiro module $H_{R[\\delta^{-1/2}],\\xi}$, with $\\xi=\\sum_j [z_j]$ in the Bloch group and $\\Delta$ a fixed multiple of the discriminant. The constant term of the series at $\\zeta_m$ is $\\varepsilon_m(\\xi)^{1/m}$ times an integral element, where $\\varepsilon_m$ is the Chern-class map $K_3(K)\\to K(\\zeta_m)^\\times/(K(\\zeta_m)^\\times)^m$. A second theorem identifies two independent constructions of these series: the $q$-hypergeometric Nahm sums of Kontsevich$-$Soibelman admissible series agree with the formal Gaussian integration series of perturbative Chern$-$Simons theory. Put together, the theorems say that perturbative quantum invariants of knots and 3-manifolds are not merely $p$-adically integral but live in a module whose gluing is governed by the Frobenius endomorphism of the number field.","pith_inferences":["If the restriction to $m$ prime to $\\Delta$ is later removed, the Habiro ring would cease to be a product of integral domains, and the Frobenius gluing would then determine the series at bad primes from the good ones; checking whether the series at roots of unity sharing primes with $\\Delta$ are forced in this way is a concrete test of how far the definition can be strengthened.","The module membership suggests a stronger congruence statement than the paper proves: the ratios of $f_{A,z}$ by explicit Pochhammer generators should be integral at all primes, not only in the computed examples, and this can be verified numerically for the $5_2$ knot series at $p=5$.","The same construction has a conjectural higher-weight analogue indexed by odd $K$-groups; if it exists, it would place descendants and higher-loop perturbative invariants in analogous modules."],"forward_implications":["The constant term of $f_{A,z}(q)$ at any $m$-th root of unity with $m$ prime to $\\Delta$ lies in $R[\\zeta_m]$ (Corollary 1.10).","The symmetrised series $f_{A,z}(q)f_{A,z}(q^{-1})$ lies in the Habiro ring $H_R$, and if the Bloch-group class $\\xi$ is torsion of order $r$, then $f_{A,z}(q)^r$ lies in $H_{R[\\delta^{-1/2}]}$ (Corollary 1.11).","Because $H_R$ is a finite projective module of rank $[K:\\mathbb{Q}]$ over $H_{\\mathbb{Z}[1/\\Delta]}$, the explicit series give concrete spanning families for Habiro rings of number fields.","The identity between admissible series and formal Gaussian integration transfers arithmetic properties from Donaldson$-$Thomas theory to perturbative Chern$-$Simons invariants, so the two subjects share integrality phenomena."],"supporting_citations":[{"why":"Supplies the unit $\\varepsilon_m(\\xi)$ from the Chern class map and the integrality statement for constant terms on which the proof of Theorem 5 relies.","marker":"[10]"},{"why":"Defines the original Habiro ring and its gluing property, which the paper generalises to number fields.","marker":"[27]"},{"why":"Introduces admissible series and their factorisation into Pochhammer symbols, one of the two sources of the main series.","marker":"[36]"},{"why":"Provides the perturbative series of a knot invariant and the integrality experiments that motivate the rings and modules.","marker":"[26]"},{"why":"Defines the formal Gaussian integration series at $q=1$ for perturbative Chern$-$Simons theory.","marker":"[13]"},{"why":"Extends formal Gaussian integration to arbitrary roots of unity, defining the series $f_{A,z,m}$.","marker":"[14]"},{"why":"Gives the $p$-adic dilogarithm and its analytic continuation along Frobenius, used in the logarithmic gluing condition.","marker":"[11]"},{"why":"Identifies the $p$-adic dilogarithm with the $p$-adic regulator, linking Coleman's map to algebraic K-theory.","marker":"[7]"}],"fun_headline_variants":["Habiro ring for number fields, graded by K3","Knot invariants p-adically glue in Habiro modules","Frobenius gluing of root-of-unity series in number fields","Nahm sums and Chern-Simons series unify in K-theory","Knot invariants get arithmetic meaning via Habiro ring"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof of the main theorem imports the result that the constant term of each level-$m$ series is $\\varepsilon_m(\\xi)^{1/m}$ times an integral element from an external theorem about Bloch groups and units; if that theorem fails for some $m$ prime to $\\Delta$, or if the omitted $N>1$ case of the argument is not actually analogous, the element $f_{A,z}$ need not lie in $H_{R[\\delta^{-1/2}],\\xi}|_\\Delta$.","fun_headline_variants_meta":{"raw":{"variants":["Habiro ring for number fields, graded by K3","Knot invariants p-adically glue in Habiro modules","Frobenius gluing of root-of-unity series in number fields","Nahm sums and Chern-Simons series unify in K-theory","Knot invariants get arithmetic meaning via Habiro ring"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000479,"raw_usage":{"total_tokens":2419,"prompt_tokens":1037,"completion_tokens":1382,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":653,"completion_tokens_details":{"reasoning_tokens":1291}},"tokens_in":653,"tokens_out":1382,"duration_ms":11368,"temperature":1.0,"reasoning_tokens":1291,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T21:35:52.347593+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take the Nahm data for the $5_2$ knot, whose field is the cubic field of discriminant $-23$, and compute the level-5 series $f_{A,z,5}(x)$ to, say, 200 terms by formal Gaussian integration. Divide by the explicit Pochhammer generator formed from the order-24 root of unity described in the paper's Example 4.3; membership in the module requires every coefficient to be 5-integral, with no negative power of 5 in the denominator. A single coefficient with negative 5-adic valuation would refute the claimed inclusion.","supporting_citations":[{"cited_title":"Bloch groups, algebraic K-theory, units, and Nahm’s conjecture","cited_arxiv_id":null,"evidence_quote":"Supplies the unit $\\varepsilon_m(\\xi)$ from the Chern class map and the integrality statement for constant terms on which the proof of Theorem 5 relies."},{"cited_title":"Cyclotomic completions of polynomial rings","cited_arxiv_id":null,"evidence_quote":"Defines the original Habiro ring and its gluing property, which the paper generalises to number fields."},{"cited_title":"Cohomological Hall algebra, exponential Hodge structures and motivic Donaldson-Thomas invariants","cited_arxiv_id":null,"evidence_quote":"Introduces admissible series and their factorisation into Pochhammer symbols, one of the two sources of the main series."},{"cited_title":"Knots, perturbative series and quantum modularity","cited_arxiv_id":null,"evidence_quote":"Provides the perturbative series of a knot invariant and the integrality experiments that motivate the rings and modules."},{"cited_title":"Quantum modularity and complex Chern-Simons theory.Com- mun","cited_arxiv_id":null,"evidence_quote":"Extends formal Gaussian integration to arbitrary roots of unity, defining the series $f_{A,z,m}$."},{"cited_title":"Dilogarithms, regulators and p-adic L-functions","cited_arxiv_id":null,"evidence_quote":"Gives the $p$-adic dilogarithm and its analytic continuation along Frobenius, used in the logarithmic gluing condition."},{"cited_title":"The syntomic regulator for the K-theory of fields","cited_arxiv_id":null,"evidence_quote":"Identifies the $p$-adic dilogarithm with the $p$-adic regulator, linking Coleman's map to algebraic K-theory."}],"review_version":1}