{"id":"666ddde3-c2df-4b8f-9292-057827ecb960","arxiv_id":"2608.08511","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A B-infinity structure on an algebra gives a monoidal tensor product on the derived category of right modules, and for Hopf algebras this produces an algebraic proof of the Benson-Krause monoidal equivalence.","lead":"This paper finds a way to define a tensor product on the derived category of right modules over an A-infinity algebra when that algebra carries a B-infinity structure. It then applies the construction to Hopf algebras, proving a monoidal Koszul duality result that gives a purely algebraic route to a known equivalence for finite groups.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 5.9's monoidal equivalence rests on an unproved compactness assertion for 1=Y(H,k) in K(Inj-H); without it the Morita step identifying F|Loc(1) with D(E) is unsupported.","rationale":"The central construction is explicit and plausible, and the diagrammatic proofs of Theorem 4.1 are consistent with the stated conventions. The genuine weak point is the compactness assertion in Theorem 5.9: it is a hypothesis, not a consequence of previously cited results, and it is exactly the condition separating an abstract generator from a compact generator needed for the Morita equivalence. Without it, the restriction F|Loc(1) is not known to be triangulated equivalent to D(E), and the monoidal equivalence claim in Corollary 5.11 is unsupported. This is the same point the reader flagged. The omitted sign verification in Proposition 5.1 and the deferred verification in Theorem 4.14 are also real, but they affect details of the monoidal structure rather than its existence. I therefore keep the CONDITIONAL verdict and recommend that the compactness point be settled before acceptance.","tokens_in":37649,"tokens_out":25776,"duration_ms":309798,"concrete_test":"Settle compactness of 1 in K(Inj-H) directly. For a concrete finite-dimensional local Hopf algebra, e.g. H=k[C2], let C=Y(H,k) be the minimal injective resolution A --x--> A --x--> ... and let X = coprod_{i>=0} Sigma^i A. Compute H^0 Hom_H(C,X) and compare it with coprod_i H^0 Hom_H(C, Sigma^i A); if the natural map is not an isomorphism, compactness fails and the Morita step in Theorem 5.9 is invalid. More generally, prove or disprove that 1 lies in the thick subcategory of K(Inj-H) generated by the indecomposable injective H-modules; if it does not, the claimed triangulated equivalence cannot hold.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Theorem 5.9's restriction claim depends on the one-line assertion, immediately after Lemma 5.8, that 'the object 1 is compact' in K(Inj-H). This is the pivotal hypothesis in the 'standard derived Morita argument' used to identify F|Loc(1) with a triangulated equivalence. If 1 were not compact, F would not preserve coproducts, and since any triangle equivalence preserves coproducts, F|Loc(1) could not be an equivalence. Compactness is not a formality here: 1=Y(H,k) is an unbounded complex of finite-dimensional injectives, and in K(Inj-H) such complexes are not automatically compact; one must prove that every morphism from 1 to a coproduct factors through a finite sub-coproduct up to homotopy. The paper neither proves this nor cites a source for it. The same unproved compactness also underlies the identification of the restriction with D(E), so the monoidal equivalence claim in Theorem B and Corollary 5.11 is only as secure as this assertion.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper constructs monoidal triangulated structures on derived categories of right A-infinity modules over a B-infinity algebra. For a B-infinity algebra A, the authors define an induction functor iota from right A-infinity modules to A-infinity bimodules using the antipode of the associated dg Hopf algebra T^c(sA), and set M ⊠_A N = M ⊗^∞_A iota(N). They prove Theorem A (Theorem 4.1) that (D^r_∞(A), ⊠_A, A) is a monoidal triangulated category, with unit and associativity constraints given by explicit bimodule quasi-isomorphisms. For brace B-infinity algebras they prove a comparison theorem (Theorem 4.14) identifying iota(M) with the original dg bimodule. The main application is to finite-dimensional Hopf algebras: the Yoneda dg algebra E = Y(k,k) is shown to carry a brace B-infinity structure, and the Koszul duality functor F = Hom_H(Y(H,k), -) : K(Inj-H) -> D(E) is triangulated lax monoidal; its restriction to the localizing subcategory generated by Y(H,k) is claimed to be a monoidal triangulated equivalence (Theorem 5.9), with a monoidal equivalence in the local case (Corollary 5.11), recovering the Krause/Benson-Krause equivalence. The final section treats graded-commutative algebras, a non-local Hopf algebra where laxness is strict, and elementary 2-groups with coproduct-dependent brace operations.","tokens_in":37915,"tokens_out":9944,"duration_ms":110864,"significance":"If the proofs are completed, this paper provides a concrete A-infinity-algebraic framework for tensor products on right module categories and a purely algebraic proof of a monoidal equivalence previously obtained through classifying spaces. The paper's strengths are its explicit formulas for iota, the unit and associativity quasi-isomorphisms, and the informative examples, especially the demonstration that the monoidal structure can depend on the Hopf coproduct. The construction is not an abstract transfer along a fully faithful embedding; the authors explicitly address the non-fullness of iota. However, the load-bearing compactness assertion in the proof of Theorem 5.9 and the omitted verifications in Proposition 5.1 and Theorem 4.14 currently prevent the manuscript from being fully convincing as written.","major_comments":[{"comment":"In the proof of Theorem 5.9, immediately after Lemma 5.8, the sentence 'The object 1 is compact' is asserted without proof or citation. Compactness of 1 = Y(H,k) in K(Inj-H) is load-bearing: it is the hypothesis that makes the 'standard derived Morita argument' identify F|Loc(1) with a triangulated equivalence, and without it F need not preserve coproducts. Since 1 is an unbounded complex of finite-dimensional injectives, compactness is not automatic and must be proved. Please supply a proof or a precise reference (for instance, via the equivalence K(Inj-H) ≃ D(H) and compactness of k in D(H)). The related generation assertion used in the proof of Corollary 5.11 also needs justification.","section":"Section 5.4, proof of Theorem 5.9"},{"comment":"Proposition 5.1 asserts that E = Y(k,k) is a brace B-infinity algebra and that Y(X,Y) is a brace B-infinity module over E, but the proof ends with 'We omit the routine sign verification' for the module identities, and the algebra part is transferred from the standard brace structure on C*(H,H). The module identities are load-bearing: Theorem 5.3 and Lemma 5.8 use the brace module structure to identify iota(F(X)) with F(X), and hence the lax monoidal structure in Theorem 5.9 depends on it. The authors should write out the verification of the brace-module identities (Definition 4.11) or cite a source that covers exactly this module case.","section":"Section 5.2, Proposition 5.1"},{"comment":"The proof of Theorem 4.14 says that the verification that I_M is an A-infinity-bimodule morphism is 'the same graphical computation' as in Theorem 3.8, with the brace-module identities replacing the corresponding brace-algebra identities. These computations are not literally the same: the brace-module identities of Definition 4.11 (higher pre-Jacobi, distributivity, higher homotopy) are involved in a different configuration from the brace-algebra identities used in Theorem 3.8. Since Theorem 4.14 is the bridge between the abstract construction and the Hopf-algebra applications, the proof should indicate which axioms are used at which step, or provide the computation.","section":"Section 4.1, Theorem 4.14"}],"minor_comments":[{"comment":"Reference [3] contains the typo 'constrcutions' and reference [28] contains the typo 'strcutre'; both should be corrected.","section":"References"},{"comment":"The sign epsilon_q in (5.4) is presented without derivation; a short explanation of how it follows from the Koszul convention stated in the introduction would improve reproducibility.","section":"Section 5.2, formula (5.4)"},{"comment":"The symbol 1 is used both for the unit object of the monoidal structure and for the injective resolution Y(H,k); given how central this object is, a distinct notation such as ℒ or Ι might reduce ambiguity.","section":"Section 5.3"}],"recommendation":"major_revision","confidential_remarks":"The stress-test concern about compactness is valid: the proof of Theorem 5.9 asserts compactness of Y(H,k) in K(Inj-H) without proof or citation. I believe the assertion is likely true and fixable via standard equivalences, but it is load-bearing and must be addressed. The omitted brace-module verification in Proposition 5.1 is also worth requiring. The paper is substantial and likely correct, but these gaps justify a major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The main construction here is real. The idea of using the antipode of the B-innity-Hopf structure to induce a bimodule iota(N) from a right module N, and then defining M box.N = M otimes^infty_A iota(N), is a genuine new mechanism. The explicit quasi-isomorphisms iota(A) equiv A and iota(M) otimes iota(N) equiv iota(M box.N) are the right technical core, and the brace-B-innity-module comparison in Theorem 4.14 is a useful bridge. The examples in Section 6, especially the dependence on the coproduct, are informative and correctly show that the monoidal structure remembers the Hopf structure.\n\nThe soft spots are in the Hopf application. The proof of Theorem 5.9 asserts, without proof or citation, that the injective resolution 1=Y(H,k) is compact in K(Inj-H). This is load-bearing: the standard derived Morita argument needs compactness to turn the restriction of F to Loc(1) into an equivalence onto D(E). In K(Inj-H), compact objects are not the unbounded complexes of finite-dimensional injectives; the Yoneda resolution is precisely such a complex, and compactness is not a formality. The stress-test note is right. I could not find a proof of this assertion anywhere in the paper, and the statement itself is suspicious. Without compactness, the identification of F restricted to Loc(1) with D(E) is unsupported.\n\nTwo smaller issues. Proposition 5.1 says the sign verification is omitted; in a diagram-calculus paper, signs are where errors hide, so this should be filled in. Theorem 4.14 defers the verification to an analogous computation; acceptable but worth checking. Also, I would ask the authors to clarify why F=Hom_H(1,-), with 1 an injective resolution, agrees with the usual Koszul duality functor RHom_H(k,-). For a complex of injectives X, Hom out of an injective resolution of the first argument is not automatically the derived Hom, and it is worth spelling out.\n\nThe paper overclaims in the abstract relative to what is proved: the monoidal equivalence of Theorem B is only as secure as the compactness assertion. That said, the construction of Theorem A is original and likely correct, and the paper deserves a serious referee. A referee should ask for a proof or citation for compactness, or a different argument that avoids it.","headline":"Theorem A gives a genuine new monoidal construction, but Theorem 5.9's monoidal equivalence rests on an unproved compactness assertion that the authors need to justify or replace.","tokens_in":38398,"tokens_out":33677,"would_cite":true,"duration_ms":356473,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["16E05","13D03","16G20","16E60"],"pacs":[],"model":"deepseek-v4-flash","headline":"A B∞-structure gives derived right-module categories a tensor product.","keywords":["B∞-algebra","A∞-module","derived category","monoidal triangulated category","Hopf algebra","Koszul duality","Yoneda dg algebra","brace algebra"],"falsifier":"Test compactness of $\\mathbf{1}=\\mathcal{Y}(H,\\Bbbk)$ in $K(\\mathrm{Inj}\\text{-}H)$: for a family of injective complexes $\\{I_i\\}$, check whether the natural map $\\bigoplus_i \\mathrm{Hom}_{K(\\mathrm{Inj}\\text{-}H)}(\\mathbf{1}, I_i)\\to \\mathrm{Hom}_{K(\\mathrm{Inj}\\text{-}H)}(\\mathbf{1}, \\bigoplus_i I_i)$ is an isomorphism. A single family where it is not disproves the unstated assumption used in the proof; conversely, proving it for all families in a non-local example would show the monoidal equivalence extends beyond the local case.","tokens_in":37456,"feed_emoji":"🧮","tokens_out":8894,"duration_ms":81985,"temperature":0.7,"pith_summary":"An ordinary noncommutative algebra has no tensor product on right modules alone, since tensoring over the algebra needs a compatible left action on the second factor. This paper proves that a $B_\\infty$-structure—an $A_\\infty$-algebra whose bar construction is a dg bialgebra—repairs exactly that missing action: the antipode of the cofree dg Hopf algebra induces an induction functor $\\iota$ that turns every right $A_\\infty$-module into an $A_\\infty$-bimodule, and the formula $M\\boxtimes_A N = M\\otimes^\\infty_A \\iota(N)$ makes the derived category of right $A_\\infty$-modules a monoidal triangulated category. Applied to a finite-dimensional Hopf algebra $H$, the Yoneda dg algebra $\\mathcal{Y}(\\Bbbk,\\Bbbk)$ of the trivial module carries a brace $B_\\infty$-structure coming from the coproduct, and the Koszul duality functor from injective $H$-complexes to modules over that Yoneda algebra becomes a monoidal triangulated equivalence wherever the trivial module's injective resolution generates. A reader should care because this supplies an explicit, purely algebraic construction of a tensor structure in settings where none was available, and it gives a new proof of a monoidal equivalence previously obtained through classifying spaces.","feed_headline":"A B∞-algebra puts a tensor product on derived right-module categories","feed_subtitle":"An explicit induction functor makes Hopf Koszul duality a monoidal equivalence on the localizing subcategory.","key_machinery":"The load-bearing object is the induction functor $\\iota$ from right $A_\\infty$-modules to $A_\\infty$-bimodules, produced by the antipode of the dg Hopf algebra $T^c(sA)$ associated to a $B_\\infty$-algebra. Explicitly, $\\varphi=\\widehat{\\mu}(S\\otimes 1)$ is a dg coalgebra morphism $T^c(sA)^{\\mathrm{op}}\\otimes T^c(sA)\\to T^c(sA)$, and $\\iota$ sends a right module $(M,\\rho)$ to the bimodule whose left structure is obtained by feeding the right action through $\\varphi$; for brace $B_\\infty$-algebras a comparison theorem identifies $\\iota(M)$ with the original dg bimodule when $M$ is a brace module. These explicit quasi-isomorphisms, verified by string diagrams, carry the monoidal axioms: they supply the unit constraint, the pentagon, and the triangle identity, and in the Hopf application they identify the derived tensor product over the Yoneda algebra with the diagonal tensor product of injective $H$-complexes.","core_discovery":"The central claim is that if $A$ is a $B_\\infty$-algebra, the derived category $\\mathcal{D}^\\mathrm{r}_\\infty(A)$ of right $A_\\infty$-modules is a monoidal triangulated category with unit $A$ and tensor product $M\\boxtimes_A N = M\\otimes^\\infty_A \\iota(N)$. The functor $\\iota$ is built from the antipode $S$ of the cofree dg Hopf algebra $T^c(sA)$ via the coalgebra map $\\varphi = \\widehat{\\mu}(S\\otimes 1)$, and the hard part is that $\\iota$ is faithful but not full, so the unit and associativity constraints must be explicit $A_\\infty$-bimodule quasi-isomorphisms $\\iota(A)\\simeq A$ and $\\iota(M)\\otimes^\\infty_A \\iota(N)\\simeq \\iota(M\\otimes^\\infty_A \\iota(N))$. The paper further claims that for a finite-dimensional Hopf algebra $H$, the Yoneda dg algebra $\\mathcal{Y}(\\Bbbk,\\Bbbk)$ is a brace $B_\\infty$-algebra and the Koszul duality functor $F=\\mathrm{Hom}_H(\\mathcal{Y}(H,\\Bbbk),-)$ is triangulated lax monoidal; restricted to the localizing subcategory generated by $\\mathcal{Y}(H,\\Bbbk)$ it is a monoidal triangulated equivalence, and if $H$ is local then $F$ itself is such an equivalence.","pith_inferences":["If compactness of $\\mathcal{Y}(H,\\Bbbk)$ in $K(\\mathrm{Inj}\\text{-}H)$ holds beyond the local case, the monoidal equivalence would extend from the localizing subcategory to all injective complexes, making Koszul duality a monoidal invariant of arbitrary finite-dimensional Hopf algebras.","The examples show the product remembers the coproduct, not just the underlying algebra or the Ext-algebra; this suggests the monoidal structure can distinguish Hopf structures that ordinary cohomological invariants cannot.","The same induction construction should work for dg bialgebras or homotopy-coherent Hopf objects, transferring the monoidal structure to derived categories of modules in settings where antipodes are only defined up to homotopy; this is testable because the only ingredient needed is a dg coalgebra morphism $\\varphi$."],"forward_implications":["Every $B_\\infty$-algebra $A$ yields a monoidal triangulated derived category of right $A_\\infty$-modules, with a tensor product defined by a concrete formula rather than by abstract transfer.","For a finite-dimensional Hopf algebra $H$, the derived category of the Yoneda dg algebra $\\mathcal{D}(\\mathcal{Y}(\\Bbbk,\\Bbbk))$ is a closed monoidal triangulated category, with brace operations inherited from the coproduct.","The Koszul duality functor $K(\\mathrm{Inj}\\text{-}H)\\to \\mathcal{D}(\\mathcal{Y}(\\Bbbk,\\Bbbk))$ is triangulated and lax monoidal on all of $K(\\mathrm{Inj}\\text{-}H)$, and becomes a true monoidal equivalence on the localizing subcategory generated by the injective resolution of the trivial module.","When $H$ is local, the Koszul duality functor is a monoidal triangulated equivalence, reproducing the Benson–Krause monoidal equivalence by purely algebraic means.","For graded-commutative algebras with trivial braces the construction recovers the usual derived tensor product; for non-local Hopf algebras the laxness is real, so the localizing restriction is essential."],"supporting_citations":[{"why":"defines $B_\\infty$-algebras as dg bialgebra structures on the tensor coalgebra, the central object class.","marker":"[17, Definition 5.2]"},{"why":"supplies the $A_\\infty$-module and bimodule derived-category framework and the quasi-isomorphism/homotopy-equivalence comparison used throughout.","marker":"[22, Section 4]"},{"why":"supplies the monoidal structure and internal Hom on $A_\\infty$-bimodules, from which $\\boxtimes_A$ is assembled.","marker":"[13, Proposition 2]"},{"why":"provides the brace $B_\\infty$-structure on Hochschild cochains, which is transferred to the Yoneda dg algebra through the embedding $\\Omega$.","marker":"[16]"},{"why":"introduces the Yoneda dg category and its bar-type morphism complexes used to define $\\mathcal{Y}(\\Bbbk,\\Bbbk)$ and $\\mathcal{Y}(H,\\Bbbk)$.","marker":"[10]"},{"why":"identifies the dg endomorphism algebra of the injective resolution $\\mathcal{Y}(H,\\Bbbk)$ with the Yoneda dg algebra, a key step in the Koszul duality comparison.","marker":"[9, Proposition 3.11]"},{"why":"provides the Benson–Krause monoidal equivalence via $C^*(BG;k)$ that Corollary 5.13 compares with the new algebraic model.","marker":"[6, Theorems 4.2 and 7.8]"}],"fun_headline_variants":["B∞-algebras build monoidal structures on derived categories","Hopf Koszul duality becomes monoidal via B∞-algebras","New tensor product on derived modules from B∞-structure","B∞-algebras make derived categories monoidal triangulated"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof of Theorem 5.9 assumes without proof or citation that the injective resolution $\\mathcal{Y}(H,\\Bbbk)$ is compact in the homotopy category $K(\\mathrm{Inj}\\text{-}H)$; if this object fails to be compact, the derived Morita argument that identifies the localizing subcategory it generates with $\\mathcal{D}(\\mathcal{Y}(\\Bbbk,\\Bbbk))$ no longer applies.","fun_headline_variants_meta":{"raw":{"variants":["B∞-algebras build monoidal structures on derived categories","Hopf Koszul duality becomes monoidal via B∞-algebras","New tensor product on derived modules from B∞-structure","B∞-algebras make derived categories monoidal triangulated"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000726,"raw_usage":{"total_tokens":3449,"prompt_tokens":1334,"completion_tokens":2115,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":950,"completion_tokens_details":{"reasoning_tokens":2054}},"tokens_in":950,"tokens_out":2115,"duration_ms":16205,"temperature":1.0,"reasoning_tokens":2054,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T04:33:43.167990+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Test compactness of $\\mathbf{1}=\\mathcal{Y}(H,\\Bbbk)$ in $K(\\mathrm{Inj}\\text{-}H)$: for a family of injective complexes $\\{I_i\\}$, check whether the natural map $\\bigoplus_i \\mathrm{Hom}_{K(\\mathrm{Inj}\\text{-}H)}(\\mathbf{1}, I_i)\\to \\mathrm{Hom}_{K(\\mathrm{Inj}\\text{-}H)}(\\mathbf{1}, \\bigoplus_i I_i)$ is an isomorphism. A single family where it is not disproves the unstated assumption used in the proof; conversely, proving it for all families in a non-local example would show the monoidal equivalence extends beyond the local case.","supporting_citations":[],"review_version":1}