{"id":"92d3eb37-d165-49c9-8dee-2590099e35aa","arxiv_id":"2508.04532","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The pseudo-q-trace construction is shown to be the same, through the sewing-factorization theorem, as the categorical end over the module category, proving conjectures of Gainutdinov-Runkel and Arike-Nagatomo.","lead":"For a broad class of vertex algebras, the authors prove long-standing conjectures equating an analytic trace construction with a categorical end construction. The proof shows these two very different-looking mathematical machineries describe the same object, a useful bridge for conformal field theory.","discovery_kind":"unification","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Partial injectivity of the canonical conformal block ω is unproved here and is the linchpin of the algebra structure on bNC; if [GZ25a] only proves a weaker sewing statement, Prop. 2.10, 2.11, and Thm. 2.13 are unsupported.","rationale":"The reader's verdict identifies two external inputs: the sewing-factorization theorem of [GZ25a] and the AUF-algebra results of [GZ25b], together with the internal partial-injectivity property of ω. I agree with the reader's assessment that the result should be conditional on independent validation of these inputs. My stress-test focuses on the partial-injectivity of ω because it is the property actually used dozens of times in the body of the paper to convert equality of contractions into equality of maps. If the imported sewing theorem does not imply this injectivity on the precise W-valued conformal block spaces used here, the associativity and module-structure theorems fail before [GZ25b] is even invoked. The paper gives no self-contained proof of Rem. 1.39, and the passages where it is applied (e.g. Prop. 2.10, 2.11, Thm. 2.13) are exactly the places where a hidden analytic assumption could enter. I do not see an internal inconsistency in the argument once the partial-injectivity statement is granted; the concern is about whether the imported statement has the required strength. Therefore I would keep the reader's CONDITIONAL verdict: the paper is plausible and coherent, but its central algebraic structure is contingent on a precise, currently unverified injectivity property. No change to the reader's verdict is needed.","tokens_in":35690,"tokens_out":6726,"duration_ms":78679,"concrete_test":"Extract the exact statement of Rem. 1.39 / Thm. 1.38 from [GZ25a] and check that it implies injectivity of the sewing map on the W-valued conformal block spaces used in Prop. 2.10, i.e. on T*_{4-pointed sphere}(bNC, W, bNC) for every W ∈ Mod(V^{⊗2}) with all sewing moduli set to 1. If the statement in [GZ25a] is weaker—e.g. only existence, or only scalar-valued blocks—try to re-derive Prop. 2.10 from the stated theorem; failure to do so confirms the gap. As an independent check, in a concrete non-rational example such as symplectic fermions, compute the kernel of the map φ ↦ ω∘φ∘ω on the relevant conformal block space; a nonzero kernel would directly falsify the partial-injectivity assumption.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central construction of the paper is the associative algebra structure on bNC, and this structure is established through repeated appeals to the 'partial injectivity of the canonical conformal block ω' (Rem. 1.39). In Prop. 2.10, Prop. 2.11, and Thm. 2.13, the authors define two conformal blocks A and B, prove that their contractions with ω on both sides agree (e.g. (2.30), (2.38)–(2.39)), and then conclude A = B by 'applying twice the partial injectivity of ω.' This requires the sewing map φ ↦ ω∘φ∘ω to be injective on the specific W-valued conformal block spaces (2.2)/(2.3) for arbitrary W ∈ Mod(V^{⊗N}) with sewing moduli 1. Existence of dual fusion products only guarantees uniqueness of the fusion product up to isomorphism; it does not by itself give this injectivity. The property is not stated or proved in the present text; it is imported from [GZ25a]. If [GZ25a] proves only a weaker injectivity statement—for example, only for scalar-valued conformal blocks, or only under different sewing radii—then the associativity identity (2.28), the equality of left and right multiplication, and the module-action theorem (2.13) lack support. Since Cor. 2.22 (AUF), Thm. 2.29 (category equivalence), Thm. 2.33 (end identification), and ultimately Thm. 3.7 all build on these results, this is the most load-bearing assumption in the paper.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the end E = ∫_{M∈Mod(V)} M ⊗_C M' for an N-graded C2-cofinite vertex operator algebra V, not assumed rational or self-dual. Using the sewing-factorization theorem of [GZ25a], it endows E, denoted bNC, with an associative C-algebra structure (Cor. 2.12) and shows that bNC is an almost-unital finite-dimensional (AUF) algebra with an involution (Cor. 2.22, Thm. 2.8). It then proves a linear category equivalence Coh_L(bNC) ≅ Mod(V) (Thm. 2.29), identifies bNC with the end in Mod(V^{⊗2}) (Thm. 2.33), and identifies the space SLF(bNC) of symmetric linear functionals with a space of torus conformal blocks (Thm. 2.36). Combining the sewing-factorization isomorphism with the pseudotrace theory of [GZ25b], the paper proves the Gainutdinov–Runkel conjecture: for a projective generator G, the pseudo-q-trace gives a linear isomorphism SLF(End_V(G)^{op}) ≅ T*_{Tz,q}(V) (Thm. 3.7). The Arike–Nagatomo conjecture is derived as a corollary (Thm. 0.1).","tokens_in":36131,"tokens_out":10094,"duration_ms":114124,"significance":"If the imported inputs are valid, this is a significant contribution. It gives a categorical/geometric explanation of pseudo-q-traces in terms of ends and proves a conjecture central to the non-semisimple Verlinde formalism. The paper's own mathematical contributions — the AUF algebra structure on bNC, the category equivalence, the end identification, and the explicit identification of the composed isomorphism with the pseudo-q-trace — are clearly valuable. The proofs are generally clear, and the graphical calculus is effective; the category-equivalence proof is elegantly organized. The principal caveat is that the central theorems depend on substantive results in the companion preprints [GZ25a] and [GZ25b], and in particular on a partial-injectivity property of the canonical conformal block ω that is not proved in this paper. The final computation in Thm. 3.7 is explicit and appears correct, but its assumptions need to be made precise and verifiable.","major_comments":[{"comment":"The associativity identity (2.28), the equality of left and right multiplication, and the module-action identities (2.35)–(2.36) are all concluded by 'applying twice the partial injectivity of the canonical conformal block ω (cf. Rem. 1.39)' after establishing equality of double contractions, e.g. (2.30) and (2.38)–(2.39). This partial injectivity is not stated as a theorem or proved in the present text; it is imported from [GZ25a]. The property is required for the specific conformal-block spaces (2.2)/(2.3), for arbitrary W ∈ Mod(V^{⊗N}), with all sewing moduli equal to 1. If [GZ25a] proves only a weaker statement — for scalar-valued blocks or under different sewing radii — then (2.28), Prop. 2.11, and Thm. 2.13 lack support, and the downstream results Cor. 2.22, Thm. 2.29, Thm. 2.33, Thm. 2.36, and Thm. 3.7 inherit the failure. The authors should either prove the required injectivity i","section":"§3.2, Thm. 3.7; (3.2), (3.5)"},{"comment":"The proof of the main theorem is a composition of two external theorems: the sewing-factorization isomorphism (3.2) from [GZ25a] and the pseudotrace isomorphism (3.5) from [GZ25b, Thm. 9.4 and 10.4]. Neither is proved or independently checked here, and the manuscript does not state the exact forms of these theorems being used, including the analytic convergence assertions imported from Thm. 1.34. Since [GZ25a] and [GZ25b] are preprints, a reader cannot verify Thm. 3.7 from this paper alone. I do not regard dependence on companion papers as an error in itself, but the authors should clearly list the specific external theorems and, where feasible, isolate the genuinely new computation (the identification of the composed isomorphism with the pseudo-q-trace).","section":"§3.2, Thm. 3.7; (3.2), (3.5)"}],"minor_comments":[{"comment":"The letter N is overloaded: it denotes the natural numbers, the number of tensor factors in Mod(V^{⊗N}), and the distinguished 2-pointed sphere N of Def. 1.26. This is a recurring source of confusion, especially in Section 2. A different symbol for the sphere would help.","section":"Notation throughout"},{"comment":"In the proof of Thm. 0.1, the equivalence Mod(V)^{op} ≅ Mod(V) via contragradient modules is asserted without comment. This is standard for grading-restricted generalized modules with finite-dimensional generalized weight spaces, but it should be stated explicitly with a reference or a one-line justification.","section":"Intro, §0.4"},{"comment":"The paper relies on [GZ25a] and [GZ25b] at several load-bearing points, but the references only give arXiv numbers. It would be much easier for the reader if each invocation named the relevant theorem (e.g. [GZ25a, Thm. X] for the sewing-factorization theorem, [GZ25b, Thm. Y] for the pseudotrace isomorphism).","section":"References; §1.6, §3.2"},{"comment":"There are several typographical artifacts that should be cleaned before publication: e.g. 'appendex' in §0.6, 'nNC' for 'bNC' in the proof of Prop. 2.11, and some corrupted placeholder symbols in Remarks 1.3 and 1.8 and Section 1.5. These do not affect the mathematics but should be fixed.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The main theorem is essentially a corollary of the authors' own companion preprints, with the novelty lying in the explicit computation identifying the composition of the sewing-factorization and pseudotrace isomorphisms with the pseudo-q-trace. The most serious issue is the unproved partial-injectivity assumption on ω; if the companion papers do not contain exactly the needed statement, the algebra structure on bNC is unsupported. I would ask the authors to supply the precise external statements and, ideally, a self-contained proof of the injectivity lemma before publication. The paper is otherwise well written and likely correct conditional on those inputs."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here is my take. This is the capstone paper in Gui-Zhang's series, and it does what it claims: it proves the Gainutdinov-Runkel and Arike-Nagatomo conjectures for C2-cofinite (not necessarily rational) VOAs, and it gives a clean bridge between pseudo-q-traces and the end construction. The genuinely new work is the associative AUF algebra structure on bNC, the category equivalence Coh_L(bNC) ≅ Mod(V), and the end identification. The sewing diagrams in Sections 2.3–2.5 are elegant, and the proof that the composition of the sewing-factorization and pseudotrace isomorphisms equals the pseudo-q-trace (Theorem 3.7) is well organized. The paper is honest about what it imports.\n\nThe soft spot is the one the stress-test flags. The partial injectivity of the canonical conformal block ω, cited as Remark 1.39, is used to convert equality of contractions into equality of maps at several load-bearing points (Props. 2.10, 2.11, Thm. 2.13). The property is not stated in this text; it is imported from [GZ25a]. If [GZ25a] only proves a weaker statement, the associativity of bNC, the left/right multiplication equality, and the module action theorem lose support. That is a real dependency, not a manufactured one. That said, these authors are working through a long series and the property is quite plausible from the universal property of fusion products; it may well be proved in [GZ25a]. The paper would be stronger if it included a precise statement of the lemma or a pointer with page/section. I would not call this fatal, but a referee must verify it.\n\nA second dependency: the sewing-factorization theorem itself and the pseudotrace theorem of [GZ25b] are not proved or independently verified here. Given the paper's aim, that is acceptable as a trilogy, but it means this paper's verdict is conditional on two companion papers. Minor points: the Arike-Nagatomo attribution to [AN03] vs [AN13] is slightly unclear, and the text has a few typos.\n\nWho is this for? Specialists in logarithmic CFT and VOA representation theory. For that audience, this is a significant result, and it deserves a serious referee. My recommendation: send it to peer review, with the explicit instruction that the referee reads [GZ25a] and [GZ25b] and checks the injectivity claim and the sewing theorem.","headline":"Serious capstone paper that proves two open conjectures, but its algebra structure on the end rests on an unproved injectivity lemma imported from a companion paper.","tokens_in":36581,"tokens_out":2400,"would_cite":true,"duration_ms":27412,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["17B69","18A30","81T40"],"pacs":[],"model":"deepseek-v4-flash","headline":"Pseudo-q-traces and categorical ends describe the same linear space of torus conformal blocks.","keywords":["vertex operator algebra","C2-cofinite","conformal blocks","pseudo-q-traces","ends and coends","projective generator","logarithmic conformal field theory","symmetric linear functionals"],"falsifier":"Choose a $C_2$-cofinite vertex operator algebra for which the vacuum torus conformal block space is explicitly known, and compute the dimension of $\\mathrm{SLF}(\\mathrm{End}_{\\mathbb{V}}(G)^{\\mathrm{opp}})$ for a projective generator $G$. If the pseudo-$q$-trace map of the main theorem is not bijective, or if different projective generators give incompatible identifications of that space with the torus blocks, the central claim is false.","tokens_in":35595,"feed_emoji":"📐","tokens_out":11595,"duration_ms":113833,"temperature":0.7,"pith_summary":"The paper proves that two constructions describe the same linear space for a broad class of vertex operator algebras: the pseudo-$q$-trace, a generalized trace with grading shifts, and the categorical end, a universal object built from all modules $M$ paired with their contragredient modules $M'$. The central result is a linear isomorphism between symmetric linear functionals on the endomorphism algebra of any projective generator and the vacuum torus conformal blocks. Along the way, the paper shows that the end carries a natural associative algebra structure, that its coherent modules reproduce the full module category of the vertex operator algebra, and that two long-standing conjectures relating these objects follow. If correct, this unifies the analytic sewing approach to conformal blocks with the categorical end and coend approach in logarithmic and non-rational conformal field theory.","feed_headline":"Pseudo-q-traces match categorical ends on torus blocks","feed_subtitle":"For any C2-cofinite vertex operator algebra, pseudo-q-traces and the end over all modules are isomorphic.","key_machinery":"The load-bearing object is the end $\\mathbb{E}=\\int_{M\\in\\mathrm{Mod}(\\mathbb{V})} M\\otimes_{\\mathbb{C}} M'$, realized geometrically as the fusion product of the trivial module along the default two-pointed sphere. The argument is carried by the sewing-factorization theorem, which identifies contractions of conformal blocks with composition, and by the canonical conformal block $\\omega$, whose partial injectivity converts equality of contractions into equality of maps. These ingredients produce the associative multiplication on $\\mathbb{E}$, the idempotent decomposition making $\\mathbb{E}$ an almost unital finite-dimensional (AUF) algebra, and finally the pseudo-$q$-trace isomorphism.","core_discovery":"On the paper's own terms, the discovery is that the end $\\mathbb{E}=\\int_{M\\in\\mathrm{Mod}(\\mathbb{V})} M\\otimes_{\\mathbb{C}} M'$ -- the universal $\\mathbb{V}^{\\otimes 2}$-module into which every pair $(M,M')$ maps -- is not just a categorical bookkeeping device. It is an associative $\\mathbb{C}$-algebra, compatible with its $\\mathbb{V}^{\\otimes 2}$-module structure, and the category of coherent left modules over this algebra is linearly equivalent to $\\mathrm{Mod}(\\mathbb{V})$. Consequently the space $\\mathrm{SLF}(\\mathbb{E})$ of symmetric linear functionals on $\\mathbb{E}$ is linearly isomorphic to the space of vacuum torus conformal blocks. Composing this with pseudotrace theory for almos","pith_inferences":["Different choices of projective generator $G$ should yield compatible identifications of $\\mathrm{SLF}$ with the same torus block space; checking this compatibility is a direct test of naturality.","The algebra structure on the end may descend to the quotient algebras $\\mathrm{A}_n(\\mathbb{V})$ and to mode transition algebras, giving those algebraic truncations a concrete cobordism-geometric origin; the paper gestures at this but leaves it open.","If the equivalence between coherent modules over the end and $\\mathrm{Mod}(\\mathbb{V})$ is appropriately algebraic, the end could serve as a finite-dimensional invariant of the representation category, useful for distinguishing non-semisimple vertex operator algebras.","In the rational case the pseudo-$q$-trace should reduce to an ordinary trace, so the theorem should recover classical genus-one modular identifications; comparing the two on a rational example would test the boundary behaviour of the result."],"forward_implications":["The pseudo-$q$-trace construction is a special case of pseudotraces on an almost unital finite-dimensional algebra, so its analytic properties follow from a finite-dimensional algebraic theory.","Every grading-restricted generalized $\\mathbb{V}$-module carries a canonical module structure over the end algebra, and the end algebra's coherent modules form an abelian category linearly equivalent to $\\mathrm{Mod}(\\mathbb{V})$.","For any finite-dimensional algebra $A$ whose finite-dimensional modules match $\\mathrm{Mod}(\\mathbb{V})$, the space of symmetric linear functionals on $A$ is linearly isomorphic to the vacuum torus conformal blocks of $\\mathbb{V}$.","The isomorphism is explicit and does not require rationality or self-duality of the vertex operator algebra, so it applies to logarithmic and other non-semisimple models.","The main theorem provides a concrete route from the endomorphism algebra of a projective generator to the vacuum torus conformal blocks."],"supporting_citations":[{"why":"Supplies the sewing-factorization theorem with its analytic convergence; this is the input that makes the fusion product and the end identification valid.","marker":"[GZ25a]"},{"why":"Provides the pseudotrace theory for almost unital finite-dimensional algebras, giving the isomorphism from symmetric linear functionals on the endomorphism algebra to those on the end.","marker":"[GZ25b]"},{"why":"States the conjecture, proved here, that the pseudo-q-trace construction realizes the isomorphism.","marker":"[GR19]"},{"why":"Introduced the pseudo-q-trace construction and the finite-dimensional algebra conjecture that appears as a corollary.","marker":"[AN13]"},{"why":"Establishes propagation and dual fusion products, which underlie the construction of the fusion product and the conformal block calculus.","marker":"[GZ23]"},{"why":"Proves convergence of sewing conformal blocks, justifying the absolute convergence of the contractions used throughout.","marker":"[GZ24]"},{"why":"Gives the Eilenberg-Watts type identification used to identify bimodule homomorphisms with symmetric linear functionals in the finite-dimensional algebra corollary.","marker":"[FSS20]"}],"fun_headline_variants":["Categorical end becomes algebra linking pseudo-q-traces to blocks","End over all VOA modules is algebra for torus blocks","Pseudo-q-traces are symmetric functionals on end algebra","End over all modules algebra proves conjectures of Gainutdinov-Runkel","Torus conformal blocks equal symmetric functionals on end algebra"],"cache_read_input_tokens":2816,"weakest_assumption_plain":"The argument assumes the imported sewing-factorization theorem with its analytic convergence, and relies on the partial injectivity of the canonical conformal block $\\omega$ to conclude that two maps are equal whenever their contractions against $\\omega$ agree; if either of these gives way, the associative algebra structure on the end is not established.","fun_headline_variants_meta":{"raw":{"variants":["Categorical end becomes algebra linking pseudo-q-traces to blocks","End over all VOA modules is algebra for torus blocks","Pseudo-q-traces are symmetric functionals on end algebra","End over all modules algebra proves conjectures of Gainutdinov-Runkel","Torus conformal blocks equal symmetric functionals on end algebra"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000805,"raw_usage":{"total_tokens":3463,"prompt_tokens":922,"completion_tokens":2541,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":666,"completion_tokens_details":{"reasoning_tokens":2465}},"tokens_in":666,"tokens_out":2541,"duration_ms":18507,"temperature":1.0,"reasoning_tokens":2465,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T23:55:29.957321+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Choose a $C_2$-cofinite vertex operator algebra for which the vacuum torus conformal block space is explicitly known, and compute the dimension of $\\mathrm{SLF}(\\mathrm{End}_{\\mathbb{V}}(G)^{\\mathrm{opp}})$ for a projective generator $G$. If the pseudo-$q$-trace map of the main theorem is not bijective, or if different projective generators give incompatible identifications of that space with the torus blocks, the central claim is false.","supporting_citations":[],"review_version":1}