{"id":"1b7e49f0-2c33-46f6-9f78-4ff5f7a0e469","arxiv_id":"2412.18753","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A root-pair framework with cyclic invariance yields Calabi-Yau completions, a bijection with Adams graded Calabi-Yau categories of Gorenstein parameter a, and a-folded cluster categories.","lead":"This paper builds a machine that starts from a fractional root of a Serre duality bimodule and produces a Calabi-Yau dg category, the root analogue of Keller's classical Calabi-Yau completion. It then proves a classification of such categories and uses it to define a-folded cluster categories as Z/aZ-quotients of usual cluster categories.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 3.3 is independent, but the bijection Theorem 3.9 and its applications rest on unpublished Proposition 4.6, so the central classification claims are conditional.","rationale":"We agree with the reader that the dependence on Proposition 4.6 from the unpublished [14] is the principal load-bearing concern. A careful check of the proof of Theorem 3.3 shows that it proceeds through Theorem 3.5, Proposition 4.5, and Lemma 4.8, and does not call on Proposition 4.6; hence the core existence of Calabi-Yau completions for cyclically invariant root pairs is internally supported. However, the bijection Theorem 3.9, the a-Segre product Theorem 6.6, and the folded cluster category Theorem 5.2 all rely on Theorem 3.11, whose proof invokes Proposition 4.6. Since [14] is in preparation and shares two of the three authors with the present paper, the reader cannot verify this keystone. We also note Remark 3.4 defers the canonical left Calabi-Yau structure to future work; this is a framing limitation rather than a gap in the stated isomorphism. Accordingly, we endorse the CONDITIONAL verdict: the paper should be accepted only if Proposition 4.6 is either proved in a revised version or the preprint [14] is made available and confirmed. No basis was found for rejecting the central Calabi-Yau completion theorem itself.","tokens_in":49605,"tokens_out":11545,"duration_ms":93216,"concrete_test":"Obtain the in-preparation manuscript [14] and verify Proposition 4.6 in full; specifically, check that the hypotheses Γ∈per[0,1]Γ^e, Γ0=A, Γ1=M are sufficient for the canonical map T_A^L M→Γ to be a graded quasi-equivalence, and that no hidden assumptions such as properness of Γ0 or cofibrancy of Γ are needed. As a targeted alternative, supply a self-contained proof of Proposition 4.6 for the special case Γ=T_A^L U that arises in Theorem 3.11, where U is a cyclically invariant a-th root on a smooth dg category A; if this restricted proof fails or needs extra hypotheses, Theorem 3.9 and downstream results must be reformulated.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's most load-bearing gap is Proposition 4.6, stated in Section 4 and cited to the in-preparation manuscript [14] by Hanihara, Iyama, and Oppermann. It asserts that a positively graded dg category Γ with Γ0=A, Γ1=M and Γ∈per[0,1]Γ^e is graded quasi-equivalent to the derived tensor category T_A^L M. This proposition is used in the proof of Theorem 3.11 to show that every Σ satisfying RHom_{Σ^e}(Σ,Σ^e)[d+1]≃Σ_{≥a−1}(a) is a derived tensor category over its degree-0 part. Theorem 3.9, the bijection between Adams graded (d+1)-Calabi-Yau categories of Gorenstein parameter a and cyclically invariant root pairs, invokes Theorem 3.11 in its '(i)→(ii)' direction; the subsequent applications—the a-Segre product theorem (6.6), the quasi-Veronese corollary (6.8), and the folded cluster category theorem (5.2)—all rely on this bijection. None of these results is accompanied by a proof of Proposition 4.6, and the overlap in authorship with [14] means the reader cannot independently assess it. Importantly, the core Calabi-Yau completion Theorem 3.3 is proved via Theorem 3.5 and Lemma 4.8 without invoking Proposition 4.6, so the central construction itself is not undermined. The concern is therefore that the classification/framework claims are conditional on an unverified external result, not that the internal logic of Theorem 3.3 is flawed.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper develops a root-of-tau analogue of Keller's Calabi-Yau completion. For a smooth dg category A, an a-th root pair (U,P) consists of an invertible A-bimodule U with U^a ≃ RHom_{A^e}(A,A^e)[d] and a subcategory P generating per A through the objects P, P⊗U, ..., P⊗U^{a-1} with the appropriate vanishing conditions. The Calabi-Yau completion Π_{d+1}^{(1/a)}(A) is the idempotent truncation e(T_A^L U)e. The paper proves (Theorem 3.3) that Π is smooth and (d+1)-Calabi-Yau when the root is cyclically invariant, and that the full tensor algebra Σ = T_A^L U satisfies the weaker duality RHom_{Σ^e}(Σ,Σ^e)[d+1] ≃ Σ_{≥a-1} (Theorem 3.5). It then states a bijection (Theorem 3.9) between Adams positively graded (d+1)-Calabi-Yau categories of Gorenstein parameter a and cyclically invariant root pairs, and uses it to prove results on a-Segre products (Theorem 6.6), quasi-Veronese categories (Corollary 6.8), and folded cluster categories (Theorem 5.2). The final section gives explicit square-root pairs for even type A Dynkin quivers and describes the resulting dg path algebras.","tokens_in":49910,"tokens_out":7766,"duration_ms":70996,"significance":"If the results hold, this is a substantial and novel extension of Keller's deformed Calabi-Yau completions: it identifies a cyclic invariance condition under which the root completion is Calabi-Yau rather than merely twisted, and it provides a classification of Adams graded Calabi-Yau categories in terms of root pairs. The proof of Theorem 3.3 is essentially self-contained, and the paper contains valuable concrete computations, notably the detailed verification in Example 3.12 and the explicit Dynkin quiver constructions in Section 8. The applications to a-Segre products and folded cluster categories are natural and appear genuinely new. The main weakness is that the classification/framework statements Theorem 3.9 and Theorem 3.11 depend on Proposition 4.6, which is imported from the unpublished manuscript [14] and is not proved here. Thus the central construction is independent, but the bijection and its applications are conditional on an external result that the reader cannot currently verify. No machine-checked proofs or reproducibility artifacts accompany the paper; its checkable content consists of traditional proofs and explicit algebraic computations.","major_comments":[{"comment":"Proposition 4.6 is stated with attribution to the unpublished manuscript [14] and asserts that a positively graded dg category Γ with Γ0 = A and Γ1 = M and Γ ∈ per_{[0,1]} Γ^e is graded quasi-equivalent to the derived tensor category T_A^L M. No proof is given in the present paper. This result is used in the proof of Theorem 3.11 (the step 'Then 4.6 applies' in the proof on pages 19–20), and Theorem 3.9 relies on Theorem 3.11 in its (i)→(ii) direction on page 23. Theorems 5.2, 6.6, and Corollary 6.8 inherit this dependence through the bijection. Since [14] is listed as in preparation and is not available to the reader, the classification and framework claims are conditional. The core Calabi-Yau completion Theorem 3.3 is proved independently through Theorems 3.5 and Lemma 4.8 and is not affected, but the paper should either include a complete proof of Proposition 4.6 or explicitly state Theorems 3.9 and 3.11 as contingent on a forthcoming external result with the precise statement quoted.","section":"Section 4, Proposition 4.6; proof of Theorem 3.11; Theorem 3.9"}],"minor_comments":[{"comment":"The remark asserts that the Calabi-Yau completion carries a canonical left Calabi-Yau structure and calls this 'the correct' refinement of Theorem 3.3, but explicitly postpones the proof to another paper. This does not affect the isomorphism statement in Theorem 3.3, but the unproved structural assertion should be either proved or clearly separated as a conjecture/forthcoming result so that it is not left as an unsupported claim in the main text.","section":"Remark 3.4"},{"comment":"The cross-references in Section 1 are inconsistent: Theorem 1.2 is labelled '(=3.1)' although the stated result is Theorem 3.3, and similar mismatches appear for the references to 3.5 and 3.11. Please correct the numbering labels.","section":"Introduction, Theorem numbering"},{"comment":"The word 'Goresntein' should be 'Gorenstein' in the phrase 'positively graded (d + 1)-Calabi-Yau (non-dg) algebras of Goresntein parameter 1'. There is also a typo in the abstract ('represe ntation').","section":"Section 1.2, page 3"},{"comment":"The displayed formula in the proof of Theorem 3.3 contains the object RHom_{Σ^e}(Σ,Σ⊗Σ); this appears to be a typo and should presumably be RHom_{Σ^e}(Σ,Σ^e), the bimodule dual of Σ.","section":"Proof of Theorem 3.3, page 20"},{"comment":"Definition 1.1 is phrased for dg algebras using add A and an idempotent e, while Section 2 works with dg categories and full subcategories P; the terminology should be harmonized so that the reader can move between the two formulations without ambiguity.","section":"Definitions 1.1 and 2.5"}],"recommendation":"major_revision","confidential_remarks":"The most serious issue is the dependence of the bijection theorems on the unpublished Proposition 4.6. Since [14] is an in-preparation manuscript by the same research group, the present paper should not rest its central classification claims on it without either a full proof or a very detailed statement/sketch. I recommend major revision rather than rejection because the independent proof of Theorem 3.3 and the concrete examples are strong and valuable; the conditional part can in principle be fixed within the manuscript's scope by adding the missing material."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The core of this paper is worth taking seriously. The author defines a-th root pairs and a root analogue of Keller's Calabi-Yau completion, proves that cyclically invariant root pairs produce Calabi-Yau dg categories (Theorem 3.3), and develops a bijection between Adams graded Calabi-Yau categories of Gorenstein parameter a and such root pairs (Theorem 3.9). The examples are meaningful, not decorative: Example 3.12 genuinely shows that cyclic invariance is an extra condition and not just a restatement of the Calabi-Yau property, and Section 8 gives explicit dg path algebras for type A_{2n} that should be useful to people working on cluster categories. The proofs are detailed and internally consistent, and the paper is honest about what it defers.\n\nThe main soft spot is exactly what the stress-test note identifies. Proposition 4.6, cited to the unpublished manuscript [14] by the author with Iyama and Oppermann, is load-bearing for Theorem 3.11 and therefore for Theorem 3.9 and the applications that depend on the bijection. The reader cannot check this result independently, and the overlap in authorship makes the dependence more awkward. Importantly, Theorem 3.3 is proved without Proposition 4.6, so the central Calabi-Yau completion result itself does not share this vulnerability. The other deferred point, Remark 3.4, postpones the canonical left Calabi-Yau structure to future work; that is a minor gap for the stated theorems, not a fatal one.\n\nI would send this to a serious referee. The right referee will know the dg category literature and can assess whether Proposition 4.6 is plausible and whether the dependence can be removed or made explicit. The paper should not be accepted as-is without either a proof of Proposition 4.6, or a clear statement that Theorem 3.9 and its applications are conditional on the companion manuscript. But it is a real contribution with a new construction, useful examples, and a credible main theorem, so it deserves referee time.\n\nWho gets value: representation theorists working on Calabi-Yau categories, dg categories, and cluster categories. This is not a paper for a general audience, but for the right reader it is substantial and worth engaging with.","headline":"Solid root-of-tau Calabi-Yau completion with an independent core theorem; the classification results rest on an unpublished proposition.","tokens_in":50479,"tokens_out":1519,"would_cite":true,"duration_ms":16709,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["16E45","18G80","16E35","16G10","16S38","14A22"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves that a cyclically invariant $a$-th root pair on a smooth dg category gives a smooth $(d+1)$-Calabi-Yau completion, and that the construction is a bijection onto Adams graded Calabi-Yau dg categories of Gorenstein…","keywords":["Inverse dualizing bimodule","Root pair","Calabi-Yau completion","Graded Calabi-Yau dg category","a-Segre product","a-folded cluster category","Gorenstein parameter","Tilting-bundle theorem"],"falsifier":"To test the weakest step, construct a positively graded dg category $\\Gamma$ with $\\Gamma_0=A$, $\\Gamma_1=M$, and $\\Gamma\\in\\mathrm{per}_{[0,1]}\\Gamma^e$ for which the canonical map $T_A^L M\\to\\Gamma$ is not a graded quasi-equivalence; such an example would disprove Proposition 4.6 and break the proof of the bijection.","tokens_in":49317,"feed_emoji":"🔄","tokens_out":11182,"duration_ms":88719,"temperature":0.7,"pith_summary":"The paper extends the Calabi-Yau completion construction to roots of shifted inverse dualizing bimodules, the bimodule counterpart of roots of shifted Serre functors. Given a smooth dg category $A$, an integer $d$, and a positive integer $a$, it defines an $a$-th root pair $(U,P)$: $U$ is an invertible bimodule with $U^{\\otimes_A a}\\simeq \\mathrm{RHom}_{A^e}(A,A^e)[d]$, and $P$ generates $\\mathrm{per}\\,A$ through a semi-orthogonal decomposition. It builds the completion $\\Pi_{d+1}^{(1/a)}(A)=e(T_A^L U)e$ and proves that if $U$ is $a$-cyclically invariant then $\\Pi$ is smooth and $(d+1)$-Calabi-Yau; without cyclic invariance it is only twisted Calabi-Yau. The same construction yields a bijection between such completions and Adams positively graded Calabi-Yau dg categories of Gorenstein parameter $a$, and gives root versions of cluster categories and of Segre products.","feed_headline":"Roots of inverse dualizing bimodules get Calabi-Yau completions","feed_subtitle":"Cyclically invariant root pairs yield Calabi-Yau dg categories and classify Gorenstein parameter a.","key_machinery":"The load-bearing object is the $a$-th root pair $(U,P)$, together with the derived tensor category $\\Sigma=T_A^L U$ and its idempotent truncation $\\Pi=e\\Sigma e$. A root pair makes $U$ an $a$-th root of the shifted inverse dualizing bimodule and gives $\\mathrm{per}\\,A$ a rectangular semi-orthogonal decomposition with layers $P\\otimes U^i$. The mechanism that turns this data into the Calabi-Yau property is cyclic invariance: the chosen quasi-isomorphism $\\varphi\\colon \\mathrm{RHom}_{A^e}(A,A^e)\\to U^a$ of degree $-d$ is required to be stable under the cyclic permutation of the factors of $U^a\\otimes_{A^e}A$. The proof translates this into a compatibility condition between $\\varphi$ and the bimodule dual of the standard triangle for $\\Sigma$, using Casimir elements in Hochschild homology; compatibility gives the isomorphism $\\mathrm{RHom}_{\\Sigma^e}(\\Sigma,\\Sigma^e)[d+1]\\simeq\\Sigma^{\\ge a-1}$ that restricts to the Calabi-Yau isomorphism for $\\Pi$.","core_discovery":"The central discovery is that the obstruction for a root-of-tau completion to be Calabi-Yau is precisely cyclic invariance of the root. Concretely, let $A$ be smooth and let $(U,P)$ be an $a$-th root pair of $\\mathrm{RHom}_{A^e}(A,A^e)[d]$; the pair means that $U^a$ realises the shifted inverse dualizing bimodule and that $\\mathrm{per}\\,A$ has the semi-orthogonal decomposition $\\mathrm{thick}(P\\otimes U^{a-1})\\perp\\cdots\\perp \\mathrm{thick}(P)$. If the closed morphism $\\mathrm{RHom}_{A^e}(A,A^e)\\to U^a$ of degree $-d$ is stable under the cyclic $\\mathbb{Z}/a$-action, then $\\Pi_{d+1}^{(1/a)}(A)$ is smooth and satisfies $\\mathrm{RHom}_{\\Pi^e}(\\Pi,\\Pi^e)[d+1]\\simeq \\Pi$ in $D(\\Pi^e)$. Conversely, up to graded quasi-equivalence, every Adams positively graded $(d+1)$-Calabi-Yau dg category of Gorenstein parameter $a$ whose degree-zero part is perfect on both sides arises as such a completion, so cyclically invariant root pairs classify these categories.","pith_inferences":["The paper leaves implicit that the same root pair data with the same $U$ but different projective layer $P$ could produce derived-equivalent completions; Question 2.13 suggests a gauge-like redundancy in the classification that would be worth testing.","One consequence not developed in the paper is that the failure of cyclic invariance is exactly the appearance of a twist: Example 3.12 already shows the completion becomes a skew polynomial ring, so the theorem delineates a sharp boundary between Calabi-Yau and twisted Calabi-Yau behaviour.","A testable extension is to run the same bijection for root pairs of different $a$ over the same category: tensor products of roots allow mixing dimensions while keeping $a$ fixed, which should yield a rich family of Calabi-Yau categories whose Gorenstein parameter is preserved.","Should Proposition 4.6 from the in-preparation source [14] be supplied and then fail in a particular graded example, the classification Theorem 3.9 would lose its converse direction; checking its validity in the non-tensor examples of Appendix A would turn the bijection into a theorem with a complete proof."],"forward_implications":["Every cyclically invariant $a$-th root pair produces a smooth $(d+1)$-Calabi-Yau dg category, and the construction determines the Adams shift, i.e. the Gorenstein parameter $a$.","Adams graded Calabi-Yau dg categories of Gorenstein parameter $a$ are classified up to graded quasi-equivalence by cyclically invariant root pairs, generalizing the classical degree-one bijection.","The $a$-Segre product of a $(d+1)$-Calabi-Yau dg category and an $(e+1)$-Calabi-Yau dg category of the same Gorenstein parameter $a$ is again Calabi-Yau of dimension $d+e+1$; in particular quasi-Veronese subcategories of these categories remain Calabi-Yau.","The cluster category of the completion is the triangulated hull of the orbit category $\\mathrm{per}\\,A/-\\otimes^L_A U$, a $\\mathbb{Z}/a\\mathbb{Z}$-quotient of the usual cluster category, and it carries a $d$-cluster tilting object.","For cyclic root pairs over Dynkin quivers of type $A_{2n}$, the completion has explicit dg path algebra presentations, giving concrete examples of folded cluster categories."],"supporting_citations":[{"why":"the deformed Calabi-Yau completion construction whose a=1 case this paper adapts to roots","marker":"[23]"},{"why":"supplies the Casimir-element and cyclic-invariance technique used in Proposition 4.5 and Theorem 3.3","marker":"[24]"},{"why":"states Proposition 4.6 on graded quasi-equivalence to a derived tensor category, used in Theorem 3.11 and the bijection Theorem 3.9","marker":"[14]"},{"why":"defines d-representation infinite algebras and the a=1 correspondence that Corollary 3.10 generalizes","marker":"[15]"},{"why":"gives the cluster category construction and its Calabi-Yau and cluster-tilting properties used in Theorem 5.2","marker":"[1]"},{"why":"supplies triangulated hulls of orbit categories, used to describe the a-folded cluster category","marker":"[21]"},{"why":"provides the Adams-graded dg category and Gorenstein-parameter framework for Sections 3 and the appendix","marker":"[13]"},{"why":"introduces a-Segre products and their non-commutative resolutions, basis for Section 6","marker":"[12]"}],"fun_headline_variants":["Cyclic invariance makes root completions Calabi-Yau","Cyclic symmetry: the key to Calabi-Yau completions","Root pairs with cyclic invariance give Calabi-Yau","Cyclic invariance classifies Calabi-Yau root completions","Adams graded Calabi-Yau from cyclic root pairs"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is the in-preparation result [14] that a positively graded dg category $\\Gamma$ with $\\Gamma_0=A$, $\\Gamma_1=M$, and $\\Gamma\\in \\mathrm{per}_{[0,1]}\\Gamma^e$ is graded quasi-equivalent to the derived tensor category $T_A^L M$; Theorems 3.11 and the bijection 3.9 cite this without proof, and if it fails the converse correspondence collapses.","fun_headline_variants_meta":{"raw":{"variants":["Cyclic invariance makes root completions Calabi-Yau","Cyclic symmetry: the key to Calabi-Yau completions","Root pairs with cyclic invariance give Calabi-Yau","Cyclic invariance classifies Calabi-Yau root completions","Adams graded Calabi-Yau from cyclic root pairs"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000539,"raw_usage":{"total_tokens":2658,"prompt_tokens":1093,"completion_tokens":1565,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":709,"completion_tokens_details":{"reasoning_tokens":1483}},"tokens_in":709,"tokens_out":1565,"duration_ms":11106,"temperature":1.0,"reasoning_tokens":1483,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T04:30:21.602038+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"To test the weakest step, construct a positively graded dg category $\\Gamma$ with $\\Gamma_0=A$, $\\Gamma_1=M$, and $\\Gamma\\in\\mathrm{per}_{[0,1]}\\Gamma^e$ for which the canonical map $T_A^L M\\to\\Gamma$ is not a graded quasi-equivalence; such an example would disprove Proposition 4.6 and break the proof of the bijection.","supporting_citations":[{"cited_title":"Herschend, O","cited_arxiv_id":null,"evidence_quote":"defines d-representation infinite algebras and the a=1 correspondence that Corollary 3.10 generalizes"},{"cited_title":"Erratum to \"Deformed Calabi-Yau completions\"","cited_arxiv_id":"1809.01126","evidence_quote":"supplies the Casimir-element and cyclic-invariance technique used in Proposition 4.5 and Theorem 3.3"},{"cited_title":"Hanihara, O","cited_arxiv_id":null,"evidence_quote":"states Proposition 4.6 on graded quasi-equivalence to a derived tensor category, used in Theorem 3.11 and the bijection Theorem 3.9"}],"review_version":1}