{"id":"68d2658a-a978-4c9a-adf3-b8d8a2659fc0","arxiv_id":"2411.19595","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The paper establishes A-infinity equivalences between curved infinity-local systems, projectively flat graded vector bundles, and curved loop-space representations, recovering twisted sheaves via a Riemann-Hilbert correspondence.","lead":"This paper proves a curved, twisted version of the higher Riemann-Hilbert correspondence: for any fixed closed 2-form h, several descriptions of h-twisted local systems (curved infinity-local systems, projectively flat graded connections, and curved loop-space representations) are all equivalent as A-infinity categories. It also shows these categories match twisted sheaves, and that over the real numbers projectively flat bundles are actually flat.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Non-compact Dolbeault case delegates crucial Assumption 6/7 to a compact-manifold lemma without verification; Theorem 7.4.3's advertised generalization hangs on it.","rationale":"The paper's headline result includes the Dolbeault application to possibly non-compact complex manifolds (abstract, Theorem 7.4.3). The equivalence there depends on Assumption 6/7, which supplies the local finite projective model needed for J to be essentially surjective and to land in DB_perf. Section 7.4 dismisses this with a citation to Block's compact-manifold lemma. Because the paper neither reproduces the lemma nor verifies its compactness-related hypotheses, this is the weakest point in the chain of the paper's advertised generalization. I agree with the reader's identification. A direct check of Block's lemma and a non-compact test case (e.g., C^2 with nonzero h) would settle whether Theorem 7.4.3 holds as stated. If it fails, the main Sections 4–5 results could still be correct, but the Dolbeault application and the abstract's claim would need revision. If it holds, the conditional verdict can be upgraded to accept after adding the proof. The other concerns (Lemma 4.5.4's deferred verification, typos in Theorem 3.3.9 and 7.4.3, the corrupted reference in Section 6.1) are less load-bearing, since Lemma 4.5.4 is a straightforward verification and the typos are editorial. Thus the reader's CONDITIONAL verdict is appropriate, and no change is recommended.","tokens_in":71291,"tokens_out":14589,"duration_ms":126598,"concrete_test":"Inspect [Blo09, Lemma 4.1.5] and identify every place where compactness of X is used in its statement or proof. Then test the non-compact assertion directly on X = C^2 with a non-zero closed (0,2)-form (e.g., h = e^{-|z_1|^2-|z_2|^2} d\\bar z_1 ∧ d\\bar z_2 as a smooth form): for the trivial cohesive module (Ω^{0,*}, \\bar∂ + h^0) with a chosen h^0 satisfying \\bar∂ h^0 = h, determine whether its cohomology sheaf is locally coherent and locally free over O_X; if it is, write out the adaptation of Block's lemma to a Stein cover of a general non-compact complex manifold; if it is not, construct a concrete counterexample to Assumption 6/7, which would falsify Theorem 7.4.3.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 7.2's Assumption 6 (referred to as 'Assumption 7' in Sections 7.3–7.4) requires that every local cohesive module be homotopy equivalent to one of the form A^*⊗_R W^* with W^* a bounded, finitely generated, projective graded R-module. This is the step that makes J land in globally bounded perfect twisted sheaves and yields essential surjectivity (Prop. 7.2.18). In the de Rham case the assumption is immediate. In the Dolbeault case, Section 7.4 states 'Assumption 7 follows by [Blo09, Lemma 4.1.5]' without reproducing the lemma or checking its hypotheses. Block's original statement was for compact complex manifolds, and it is not obvious that his proof carries over to possibly non-compact X; it could use finite-dimensionality of cohomology, a compact exhaustion, or other compactness-dependent arguments. The abstract explicitly advertises the non-compact Dolbeault generalization, so a failure of Assumption 6/7 would collapse Theorem 7.4.3 even though the Sections 4–5 loop-space equivalences might remain intact. The paper should either prove the assumption directly or verify that Block's lemma applies verbatim to non-compact manifolds.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a curved version of the higher Riemann-Hilbert correspondence. For a fixed closed 2-form h on a smooth manifold M, it introduces dg-categories of h-curved cohesive modules P(M)_∞[h], h-curved ∞-local systems Loc(M)_∞[H], and curved representations of the singular simplicial set of the Moore loop space, and shows via A∞-quasi equivalences that they are all equivalent (Theorems 4.5.6, 5.2.11, and 5.4.5). It then proves 1-categorical reductions to projectively flat vector bundles and projective representations of π_1(M), and in the final section gives a general twisted-sheaf correspondence: cohesive modules over a curved sheaf of dg-algebras are equivalent to globally bounded perfect twisted sheaves (Theorem 7.2.20), with applications to the de Rham algebra (Theorem 7.3.3) and to the Dolbeault algebra of a possibly non-compact complex manifold (Theorem 7.4.3), generalizing a result of Block.","tokens_in":71364,"tokens_out":12291,"duration_ms":99138,"significance":"If the main results hold, they provide a coherent dg-enhancement framework for the bounded derived category of twisted locally constant sheaves and a useful categorical description of projectively flat graded bundles. The paper's technical core is substantial: the holonomy-form machinery is developed in detail, the A∞-relations are verified explicitly, and the twisted-complex criterion of Theorem 3.3.9 is a useful general tool. The h=0 specialization correctly matches the established higher Riemann-Hilbert results of [BS14], [Hol14a], [Hol14b], [CHL21], and [AS16], which provides a consistency check. The main advertised new applications, especially the non-compact Dolbeault theorem and the projectively flat loop-space equivalence, are significant if the deferred assumptions are verified.","major_comments":[{"comment":"Assumption 6 (referred to as 'Assumption 7' in §7.3 and §7.4) is load-bearing for the Dolbeault theorem, but it is not verified for non-compact complex manifolds. The assertion in §7.4 that 'Assumption 7 follows by [Blo09, Lemma 4.1.5]' is insufficient: Block's lemma is formulated for compact complex manifolds, and the manuscript does not reproduce the lemma or check that its hypotheses continue to hold when X is non-compact, even though the abstract advertises exactly that generalization. Since Proposition 7.2.18 and hence Theorem 7.4.3 depend on this local realizability property, the paper should either prove Assumption 6 directly for non-compact X or identify the precise hypotheses in [Blo09] that apply. The same issue affects §7.3, where Assumption 7 is said to follow from homotopy invariance without a demonstration.","section":"§7.2–7.4"},{"comment":"The proof of the isomorphism of complexes C^*(M,(Q,∇)) ≅ Hom^*_{Loc(M)^0[H]}(RH(E),RH(F)) is omitted ('we leave it to the reader'). This lemma is the mechanism for quasi-fully faithfulness of RH^0, and Theorem 4.5.6 depends on it. Please provide a complete proof or a precise reference covering the curved case; in particular, the identity relating parallel transport and the exponential factors, and the comparison of the two differentials, should be written out.","section":"Lemma 4.5.4"},{"comment":"Essential surjectivity of Tw(F) in Theorem 3.3.9 relies on Proposition A.9 of [AS16], which is neither stated nor proved in the manuscript. The hypotheses to be checked include the existence of the Maurer-Cartan element F, the invertible element g, and the compatibility of Tw(F) with the auxiliary-degree filtration. Please state the external proposition and verify its hypotheses in the present setting, or give a self-contained argument.","section":"Proposition 3.3.8"},{"comment":"The cohomology table for M = RP^2 × S^1 appears inconsistent with the Künneth theorem: for example H^3(RP^2 × S^1; Z) = 0 and H^2(RP^2 × S^1; C) = 0, while the table lists Z/2 × Z and Z/2 × C respectively. Since the example is used to show that the projective representation f does not arise from a projectively flat vector bundle, the computation of the image of H^2(π_1(M); C^*) in H^3(M; Z) must be redone.","section":"Example 6.2.2"}],"minor_comments":[{"comment":"The statement says 'Tw(F) : Tw(C) → Tw(C)'; the target should be Tw(D).","section":"Theorem 3.3.9"},{"comment":"The assumptions are numbered 1–6, but §7.3 and §7.4 refer to 'Assumption 7'; please renumber or cross-reference consistently.","section":"§7.2"},{"comment":"The displayed equivalence repeats the same category on both sides: H^0((Ω^{0,*}(X),h)-Mod_coh) → H^0((Ω^{0,*}(X),h)-Mod_coh); one side should presumably be the sheafified category (A^*,h)-Mod_coh from §7.2, and the variable in h ∈ Ω^{0,2}_{cl}(M) should be X.","section":"Theorem 7.4.3"},{"comment":"The phrase 'By Definition 2 in Proposition 4.2.2' should refer to property (2) of that proposition, and the references to property 4 later in the proof should be made precise.","section":"Lemma 4.6.4"},{"comment":"The table in Example 6.2.1 and the surrounding notation for cohomology groups would benefit from explicit coefficient conventions, since singular cohomology with coefficients in C^* and C are used side by side.","section":"§6.2"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is very long and appears to be thesis-derived; the original core in Sections 4–5 is substantial and likely correct, but verification is made harder by the number of deferred arguments. The main risk is the unverified non-compact Dolbeault assumption and the omitted proofs of Lemma 4.5.4 and the cited Proposition A.9 of [AS16]; these are fixable in revision. No concerns about attribution are warranted: the curved statements specialize at h=0 to previously published results, and the use of the advisor's work is limited to established external benchmarks."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a real contribution, not a sketch. The A-infinity quasi-equivalences between h-curved cohesive modules, curved infinity-local systems, and curved loop-space representations (Theorems 4.5.6, 5.2.11, 5.4.5) are genuinely new, and the holonomy-form technology (defining holonomy as solutions to ODEs instead of iterated integral series, plus the smooth admissible cube-to-simplex maps) is a real improvement in presentation. The reduction via Theorem 3.3.9, showing it suffices to check quasi-equivalence on objects concentrated in degree 0, is a useful tool in its own right. The twisted-sheaf consequences in Sections 7.2 and 7.3 are well motivated, and the de Rham case works out cleanly.\n\nThe soft spots are real but not fatal, and they sit where the reader's report puts them. Lemma 4.5.4 is load-bearing for quasi-fully faithfulness in the loop-space comparison, and it is deferred with 'we leave it to the reader.' That is too important a step to leave as an exercise in a paper of this size. Section 6.1 contains a visibly corrupted passage about Block and Daenzer that needs to be fixed before publication. The typos in Theorems 3.3.9 and 7.4.3 are minor but should be caught in copyediting.\n\nThe bigger question is Assumption 6/7 in Section 7.2. I think the stress-test note is right that the Dolbeault generalization is the weak link. The paper states that Assumption 7 follows from [Blo09, Lemma 4.1.5] in the non-compact case without reproducing the lemma or checking whether Block's proof uses compactness. Block's original result was for compact complex manifolds, and the abstract advertises the non-compact generalization. If that lifting property fails away from compactness, Theorem 7.4.3 collapses, even though the Sections 4-5 equivalences would remain intact. This should be fixed either by proving the assumption directly or by verifying that Block's lemma applies verbatim to non-compact manifolds.\n\nNone of this undermines the central argument as far as I can tell. The uncurved cases specialize correctly to [BS14], [CHL21], [AS16], and Holstein's work, and the new curved statements are stated precisely. The citation pattern is fine; building on the advisor's prior work is not a problem when those results are external and published, and the new theorems do not merely repackage them.\n\nThis paper is for people working in higher Riemann-Hilbert correspondence, dg-categories, and twisted sheaves. It deserves a serious referee. Send it out, but with a request that the author prove Lemma 4.5.4, clean up Assumption 6/7 for the Dolbeault case, and fix the corrupted text. It is close to publishable in a good journal after those revisions.","headline":"A substantial, technically serious extension of the higher Riemann-Hilbert correspondence to curved/twisted settings; the main equivalences look right, but the Dolbeault non-compact generalization and a few deferred proofs need attention before I'd trust it fully.","tokens_in":72078,"tokens_out":2503,"would_cite":true,"duration_ms":20857,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["55N30","14F08","32L10","53C05","58A12"],"pacs":[],"model":"deepseek-v4-flash","headline":"The higher Riemann–Hilbert correspondence survives a fixed scalar curvature: for any closed 2-form h, curved local systems, projectively flat graded vector bundles, and curved loop-space representations form equivalent dg-categories.","keywords":["curved ∞-local systems","projectively flat graded vector bundles","higher Riemann-Hilbert correspondence","twisted sheaves","Moore loop space","cohesive modules","curved Dolbeault algebra","A∞-quasi equivalence"],"falsifier":"Find a non-compact complex manifold X and a closed (0,2)-form h such that some object of $H^{0}$(($Ω^{{0,*}}$(X),h)-Mod_coh) is not, on any neighbourhood, homotopy equivalent to $Ω^{{0,*}}$⊗_{O_X} W^* with W^* bounded and finitely generated projective over O_X; that would make J fail to land in DB_perf(X)_h and break Theorem 7.4.3, while leaving the loop-space equivalences of Sections 4–5 intact.","tokens_in":70898,"feed_emoji":"🔗","tokens_out":9479,"duration_ms":76138,"temperature":0.7,"pith_summary":"This paper extends the higher Riemann–Hilbert correspondence to settings with a fixed scalar curvature, namely a closed 2-form h on a smooth manifold M. It proves that three dg-categories—curved ∞-local systems, graded vector bundles with projectively flat Z-graded connections, and curved representations of the based Moore loop space—are all A∞-quasi equivalent for the same h. That equivalence yields dg-enhancements of the bounded derived category of twisted locally constant sheaves with finite-dimensional fibers, so curvature h is implemented by twisting sheaves by a gerbe. In the ungraded case the same framework gives a precise correspondence between projectively flat vector bundles and projective representations of the fundamental group, with a computable obstruction deciding which projective representations actually arise. A final application transfers the machinery to the curved Dolbeault algebra of a possibly non-compact complex manifold, recovering a twisted-sheaf statement previously known only in the compact case.","feed_headline":"A fixed curvature survives the higher Riemann–Hilbert correspondence","feed_subtitle":"Three dg-categories become equivalent models of twisted locally constant sheaves for any closed 2-form h","key_machinery":"The load-bearing mechanism is a family of holonomy forms, defined as unique solutions of first-order differential equations on path space rather than by summing iterated integrals. These forms assemble the parallel transport of a projectively flat graded connection into simplicial data, and the higher Riemann–Hilbert functor is built by integrating them over a collection of smooth cube-to-simplex maps satisfying explicit admissibility axioms. Around this, the paper develops a general criterion, Theorem 3.3.9, that upgrades an A∞-quasi equivalence on objects with zero curvature to an A∞-quasi equivalence of the associated dg-categories of twisted complexes, provided both curved categories are sufficiently Maurer–Cartan and split. That criterion is what lets curvature be carried through the correspondence without re-doing the full homotopy theory at each step.","core_discovery":"The paper's central claim is that scalar curvature is compatible with every level of the Riemann–Hilbert correspondence. For any closed 2-form h, Theorems 4.5.6 and 5.2.11 give A∞-quasi equivalences among the dg-category of h-curved cohesive modules, the dg-category of h-curved ∞-local systems, and the dg-category of curved representations of the singular simplicial set of the based Moore loop space. Theorem 5.4.5 then packages all scalar curvatures into an ordinary equivalence between projectively flat graded vector bundles and logarithmic projective representations of the loop space; its ungraded specialization, Theorem 6.1.9, says projectively flat bundles correspond exactly to projective representations of π1(M;x0) whose curvature class is killed by $H^{2}$(π1(M),C*) → $H^{3}$(M,Z). In Section 7 the paper proves that these dg-categories enhance the bounded derived category of twisted locally constant sheaves (Theorem 7.3.3) and, for a complex manifold, identifies cohesive modules over the curved Dolbeault algebra with globally bounded perfect twisted sheaves, with the embedding into bounded coherent twisted sheaves becoming an equivalence when the manifold is compact (Theorem 7.4.3).","pith_inferences":["The paper leaves implicit that the same Section 7 machine would work for any soft, exponentiable resolution of a coefficient sheaf; replacing de Rham or Dolbeault forms with another soft resolution would give a twisted Riemann–Hilbert correspondence in that geometry.","The projective categories PF∞(M) and LPRep(ΩM) identify objects up to tensor product by line bundles; a natural next step is to check whether the equivalence intertwines characteristic-class invariants, such as the projective Chern class of a bundle with the curvature class of the corresponding loop-space representation.","A sharper test of Theorem 7.4.3 outside compactness is to compute the essential image of J on a non-compact Stein manifold, where vanishing of coherent cohomology makes the bounded coherent side easier to describe; if the image strictly misses some bounded perfect twisted sheaf, the lifting assumption is genuinely needed, not automatic."],"forward_implications":["For each closed 2-form h, the homotopy category of h-curved cohesive modules is equivalent to the derived category of h-twisted locally constant sheaves with bounded, finite-dimensional cohomology, so scalar curvature is exactly a gerbe twist.","A projectively flat vector bundle over a connected manifold corresponds to a projective representation of π1(M;x0) whose curvature is annihilated by H^2(π1(M),C*) → H^3(M,Z); Example 6.2.2 shows the obstruction is non-trivial, since a projective representation of π1(RP^2 × S^1) with generator curvature is realized by no projectively flat bundle.","Over the real numbers, projective flatness is a disguised version of flatness: every non-zero projectively flat real bundle has exact curvature, so the h-curved category is equivalent to the flat one when [h]=0 and is zero otherwise (Theorem 6.3.2).","For a possibly non-compact complex manifold, cohesive modules over the curved Dolbeault algebra form a dg-enhancement of the bounded derived category of globally bounded perfect twisted sheaves, embedding fully faithfully into bounded coherent twisted sheaves and equalling it in the compact case."],"supporting_citations":[{"why":"Defines the flat higher Riemann–Hilbert equivalence between cohesive modules and ∞-local systems that the paper extends by a fixed scalar curvature.","marker":"[BS14]"},{"why":"Establishes the flat quasi-equivalence between cohesive modules and loop-space representations that Section 5 generalizes to curved representations.","marker":"[AS16]"},{"why":"Identifies cohesive modules over the de Rham algebra directly with perfect complexes, the pattern Section 7 adapts to twisted sheaves.","marker":"[CHL21]"},{"why":"Supplies the compact-manifold Dolbeault equivalence and the lemma used for the lifting assumption in the complex case.","marker":"[Blo09]"},{"why":"Produces the A∞ de Rham integration map, the template for the higher Riemann–Hilbert functor.","marker":"[Gug77]"},{"why":"Introduces holonomy of superconnections as iterated integrals, the model for the holonomy forms carrying the functors.","marker":"[Igu09]"},{"why":"Shows ∞-local systems are equivalent to representations, the uncurved counterpart of the curved ∞-local system statement.","marker":"[Hol14a]"},{"why":"Identifies ∞-local systems with perfect complexes of sheaves, the flat precursor of the twisted perfect-sheaf enhancement.","marker":"[Hol14b]"},{"why":"Provides twisted complexes for curved A∞-categories, which the paper adapts to pass from curved categories to dg-categories.","marker":"[Low07]"}],"fun_headline_variants":["Curved ∞-local systems match projectively flat bundles","Scalar curvature compatible with higher Riemann-Hilbert","Fixed curvature extends Riemann-Hilbert to twisted sheaves","Higher R-H correspondence works with any closed 2-form"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole twisted-sheaf conclusion in Section 7 rests on assuming that every locally projective graded module over the resolution algebra can locally be descended to a bounded, finitely generated, projective module over the underlying coefficient sheaf; in the complex-manifold application this lifting is asserted to follow from a lemma stated for compact manifolds, and if the lifting fails away from compactness the Dolbeault equivalence collapses even though the loop-space equivalences may survive.","fun_headline_variants_meta":{"raw":{"variants":["Curved ∞-local systems match projectively flat bundles","Scalar curvature compatible with higher Riemann-Hilbert","Fixed curvature extends Riemann-Hilbert to twisted sheaves","Higher R-H correspondence works with any closed 2-form"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00038,"raw_usage":{"total_tokens":2063,"prompt_tokens":1033,"completion_tokens":1030,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":649,"completion_tokens_details":{"reasoning_tokens":964}},"tokens_in":649,"tokens_out":1030,"duration_ms":8991,"temperature":1.0,"reasoning_tokens":964,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T06:01:50.720694+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Find a non-compact complex manifold X and a closed (0,2)-form h such that some object of $H^{0}$(($Ω^{{0,*}}$(X),h)-Mod_coh) is not, on any neighbourhood, homotopy equivalent to $Ω^{{0,*}}$⊗_{O_X} W^* with W^* bounded and finitely generated projective over O_X; that would make J fail to land in DB_perf(X)_h and break Theorem 7.4.3, while leaving the loop-space equivalences of Sections 4–5 intact.","supporting_citations":[],"review_version":1}