{"id":"8ab2864b-75b0-45d9-8ef0-a6474be4d719","arxiv_id":"2509.08478","paper_version":3,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":2.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"A survey of how link homology theories extend to 4-manifold invariants called skein lasagna modules, which can distinguish exotic smooth structures.","lead":"This paper surveys recent work turning link homology into invariants of smooth 4-manifolds, including the skein lasagna modules that can detect exotic smooth structures. It is a useful map of a rapidly developing program in low-dimensional topology and extended TQFT, written by one of its main contributors.","discovery_kind":"review","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Sweep-around invariance (condition c) is the load-bearing assumption; the survey does not reproduce its proof, so the entire extension to S^3, lasagna modules, and 4-manifold exotic detection rests on [MWW22].","rationale":"The reader correctly identifies condition (c) as the weakest assumption. The survey presents Theorem 2.4 as a known result but does not reproduce the proof, and the text itself emphasizes the difficulty of the sweep-around move. Since the paper is a survey, this reliance on a cited theorem is not a flaw in exposition, provided the citation is accurate and the limitations are disclosed. The paper does disclose the informal status of the 2-category generalization in Section 2.5 and the ongoing rigorization of the TQFT claim in Section 4. Thus the central claim is honestly presented as a synthesis of the literature, and the reader's ACCEPT verdict remains appropriate. No new objection changes that assessment.","tokens_in":14751,"tokens_out":15383,"duration_ms":137688,"concrete_test":"Independently re-derive the sweep-around proof of [MWW22, Section 2.2] in the general setting of Theorem 2.1: verify that the categorified Kauffman trick gives, for every tangle T, a chain homotopy between the two Reidemeister-3 maps in (2.1), and that the resulting homotopies are compatible with tangle composition. If this naturality fails for any T, condition (c) is not established; if it holds, the infinite family reduces to finitely many generating moves.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The single most load-bearing point is condition (c) of Theorem 2.1, exactly as the reader flags. Theorem 2.4 asserts that gl_N homology satisfies it, but the survey gives only a qualitative description of the proof: the sweep-around move is an infinite family of non-local moves indexed by tangles T, and the categorified Kauffman trick compares Reidemeister-3 chain maps. No chain-level computation or coherence data is shown. If that proof had a gap for any tangle, the extension to links in S^3 (Theorem 2.1.1) would fail, and the lasagna algebra, skein modules, and the [RW24] exotic-structure detection would all collapse at the first step. The paper's own narrative, that this step caused the delay between [Wal07] and [MWW22], confirms it is not a formality. A second, smaller gap is that conclusion 4 (the braided monoidal 2-category) is stated for arbitrary V, but Section 2.5 admits the construction is only informal for general H; the rigorous version in [MWW22, Section 6] is for KhR_N. This does not change the verdict on the survey, since both points are explicitly disclosed and the survey's purpose is synthesis, not new proof.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This survey, written for the ICBS 2025 proceedings, explains how link homology functors satisfying a small set of axioms can be promoted to much richer invariants: links in S^3, algebras for the lasagna operad, skein modules for 4-manifolds, and braided monoidal 2-categories. The main organizing statement is Theorem 2.1, which is attributed to Morrison-Walker-Wedrich [MWW22] and an extension of [MWW24, Theorem 2.1]. The author then reports (Theorem 2.4) that gl_N Khovanov-Rozansky homology satisfies the hypotheses, describes the lasagna skein module constructions and their properties, surveys handle-attachment formulas, surface invariants, and explicit computations including the recent exotic-4-manifold detection of Ren-Willis [RW24], and closes with a discussion of the conjectural chain-level versions of the theory in the framework of extended TQFTs. The paper is explicitly a survey: no new proofs are claimed, open questions are identified, and the main theorems are attributed to the cited literature.","tokens_in":15038,"tokens_out":5752,"duration_ms":54483,"significance":"If the framework described is correct, this is a valuable and timely survey. Its main strengths are transparency and attribution: every major theorem carries a precise reference to [MWW22], [MN22], [RW24], [SZ24], etc., and the paper openly flags where a construction is only informal (Section 2.5) or where a proof is deferred to the literature (Section 2.2). The survey makes the logical architecture of the skein lasagna program explicit and is a useful entry point for nonspecialists. It also clearly separates established results from conjectural programmatic claims, especially in Section 4.1. No new computational claims are made, so the soundness burden rests on the cited sources; the paper itself does a good job of indicating where that burden lies.","major_comments":[{"comment":"Theorem 2.1, conclusion 4, is stated for an arbitrary symmetric monoidal cocomplete target category V and for any link homology H satisfying conditions (a)-(c), but Section 2.5 explains that the construction of the braided monoidal 2-category is 'described only informally' except for the prototypical case H = KhR_N, where [MWW22, Section 6] is rigorous. As written, the theorem therefore overclaims: the reader cannot tell whether conclusion 4 is a theorem for all V or a conjecture. The statement should be restricted to the hypotheses under which the rigorous construction is known, or a precise citation to a proof for general V should be supplied. Since conclusion 4 is the bridge to the TQFT interpretation in Section 4, this is not merely a cosmetic issue, although it is readily fixable in revision.","section":"Section 2.5 and Theorem 2.1"}],"minor_comments":[{"comment":"The term 'locally V-enriched' is used without a definition; a one-sentence explanation, or a reference to the precise definition in [MWW22], would help the reader assess the exact categorical structure asserted.","section":"Section 2.5"},{"comment":"In the bullet on four-handles, the phrase '4-handles can also be freely removed' is ambiguous; it should say that removing a four-handle induces an isomorphism of skein modules, matching the citation to [MN22, Proposition 2.1].","section":"Section 3.2"},{"comment":"The entry 'B^3 x S^1  ⊔ 2m S^1' is hard to parse; if the intended operation is a connected sum rather than a disjoint union, the notation should be clarified.","section":"Table 2"},{"comment":"The footnote stating that portions of the article were edited using generative AI is a welcome disclosure, but the paper does not indicate which portions were affected; adding a brief specification would make the disclosure more useful.","section":"Footnote 5"}],"recommendation":"major_revision","confidential_remarks":"The survey is largely built on the author's own prior work with coauthors, which is natural for this type of exposition. The main issue is not the reliance on [MWW22] for the sweep-around proof, which is properly attributed, but the mismatch between Theorem 2.1's broad statement and the informal status of conclusion 4 for general V. This is fixable by a precise reformulation, but it should be corrected before publication. I did not find any independent-verification problem with the cited [RW24] computation, though an explicit statement that this computation has been checked or reproduced would strengthen the survey's reliability."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a survey, not a new-results paper. That is fine — it is a good survey. Wedrich maps the Morrison–Walker–Wedrich program from link homology to skein modules for 4-manifolds, with an honest account of what is proven, what is conjectural, and where the heavy lifting is done elsewhere. The one mildly new statement, Theorem 2.1 extended to arbitrary cocomplete symmetric monoidal targets plus a braided monoidal 2-category conclusion, appears without proof; the paper itself flags that the rigorous version in [MWW22] is for KhR_N and the general case is informal. That is a soft spot, but it is disclosed and it does not hurt the survey's purpose.\n\nWhat is good: the paper gives a clear structural picture of the lasagna operad, three equivalent formulations of skein modules, handle-attachment formulas, and a sample-computation table covering S^4, B^3×S^1, S^2×D^2, S^2×S^2, CP^2, and CP^2-bar. It also places braided monoidal 2-categories in the periodic table, which is helpful context. Open questions, such as finite-rankness of skein modules and the poorly controlled 1-handle formula, are labeled explicitly. The citation pattern is healthy: the core is [MWW22] and follow-ups, but independent computations by Manolescu–Neithalath, Sullivan–Zhang, Ren–Willis, Sullivan, and Teng show the program is being stress-tested by other groups.\n\nThe load-bearing assumption is condition (c) of Theorem 2.1, invariance under the sweep-around move. The survey gives a qualitative description of why the proof is hard — an infinite family of non-local moves, a categorified Kauffman trick — and credits [MWW22] for the details. It does not show the chain-level computation. That is normal for a survey, but worth saying explicitly: if that proof had a gap for some tangle, the extension to S^3, the lasagna algebra, and the 4-manifold invariants would collapse. Nothing in this paper suggests such a gap, and the stated delay from 2007 to 2022 points to genuine content rather than oversight. The stress-test note is right about where the weight sits, but the survey is also right to rest it there, given the attribution.\n\nMinor quibbles: footnote 5 says portions were edited using generative AI. The text is coherent and technically careful, so I do not see that as a substantive issue. A few references point to lecture notes or informal sources, which is acceptable for a survey.\n\nWho this is for: a graduate student or a topologist from a neighboring field who wants the current state of the skein-lasagna story in one place. Not for someone looking for a new theorem. I would send it to a serious referee, but the referee should judge it as a survey and not hold the lack of new proof against it.","headline":"A solid, honest survey of the skein-lasagna program; the real weight sits on the sweep-around move, and the paper tells you exactly where to look rather than proving it.","tokens_in":15507,"tokens_out":2770,"would_cite":true,"duration_ms":24740,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["57K18","57K41","57R56"],"pacs":[],"model":"deepseek-v4-flash","headline":"Three natural conditions lift link homology to 4-manifold invariants","keywords":["link homology","Khovanov-Rozansky homology","skein lasagna modules","4-manifold invariants","braided monoidal 2-categories","topological quantum field theory","exotic smooth structures","categorification"],"falsifier":"Take a concrete tangle T in the sweep-around move—for example a single crossing or a braid closure—and compute the chain map that Khovanov homology (gl_2) assigns to the two movies related by that move; if the two maps are not homotopic, Theorem 2.4 and the whole extension fail. Alternatively, find a link cobordism in $S^{3}$×I whose movie passes through the point at infinity and whose induced map on KhR_N is not the identity.","tokens_in":14580,"feed_emoji":"🧶","tokens_out":5265,"duration_ms":33769,"temperature":0.7,"pith_summary":"This survey argues that a link homology theory—a functor assigning chain complexes to links and chain maps to link cobordisms—becomes a much richer object once it satisfies three conditions: invariance under a full 2π rotation of $R^{3}$, monoidality under disjoint union, and invariance under the sweep-around move. Under these hypotheses the invariant extends from links in $R^{3}$ to links in $S^{3}$, defines an algebra for the lasagna operad, produces skein lasagna modules for pairs (W,L) of a smooth 4-manifold and a boundary link, and yields a braided monoidal 2-category with duals and adjoints. The motivating examples are the general linear (gl_N) link homologies, which satisfy the conditions via the sweep-around theorem proved by a categorified Kauffman trick. The payoff is that the resulting skein modules are computable along handle decompositions and are sensitive enough to distinguish exotic smooth structures on 4-manifolds, realizing part of the Crane–Frenkel program of categorification-based 4-dimensional TQFTs.","feed_headline":"Three natural conditions lift link homology to 4-manifold invariants","feed_subtitle":"The same skein machinery that computes 3-manifold invariants now tells apart smooth structures on 4-manifolds.","key_machinery":"The load-bearing mechanism is the sweep-around move, an infinite family of non-local cobordism moves that arise when an isotopy of a link or cobordism in $S^{3}$ passes through the point at infinity of $R^{3}$. The paper's proof that gl_N homology is invariant under this move uses a categorified Kauffman trick: it compares the chain maps assigned to Reidemeister-3 moves in the two configurations where the closing strand passes in front or behind a tangle T, exploiting the interaction between the skein relation and Reidemeister-2 and -3 moves. Once this move is controlled, the extension to $S^{3}$ is built by taking transitive systems over the groupoid of parametrizations of $R^{3}$ and over the fundamental groupoid of the link complement, and the same invariance underlies the well-definedness of lasagna operad composition, skein modules, and the braided monoidal 2-category.","core_discovery":"The central discovery is Theorem 2.1: any link homology functor H from Links($R^{3}$) to a symmetric monoidal cocomplete category V that is invariant under the trace of the 2π rotation, laxly monoidal under disjoint union, and invariant under the sweep-around move extends canonically to link homology for links in $S^{3}$, an algebra for the lasagna operad, a (4+ε)-dimensional TQFT whose top layer consists of skein modules for pairs (W,L), and a locally V-enriched braided monoidal 2-category with duals for objects and adjoints for 1-morphisms. The paper further reports that the gl_N link homologies satisfy the hypotheses: Theorem 2.4 states that the gl_N functor assigns the identity to every sweep-around move. On this basis the survey assembles evidence that skein lasagna modules are nontrivial 4-manifold invariants: they recover link homology on $B^{4}$, behave well under gluing and handle attachment, give invariants of embedded and immersed surfaces, and in [RW24] distinguish an exotic pair of knot traces by their quantum-degree −1 part.","pith_inferences":["If the sweep-around invariance were established for other link homology theories—for example link Floer homology or equivariant variants—the same Theorem 2.1 template would automatically produce new lasagna algebras and 4-manifold skein modules, without rebuilding the global framework.","The chain-level conjecture (Conjecture 4.1), if resolved with dualizable generating objects, would likely let the skein modules be promoted to a fully extended (4+ε)-TQFT in the sense of the cobordism hypothesis, with the braided monoidal 2-category upgraded to an E_2-monoidal (∞,2)-category.","The documented sensitivity to exotic structures suggests a practical test strategy: compute skein lasagna modules for other candidate exotic pairs, such as other knot traces or exotic surfaces, to probe how much of the smooth classification is captured by categorified skein theory.","Filtered or deformed skein modules (Lee-type deformations) may produce numerical invariants for 4-manifolds analogous to Rasmussen's s-invariant, extending the genus bounds already obtained for the 4-ball."],"forward_implications":["Skein lasagna modules give a computable, algebraically defined invariant of smooth 4-manifolds with boundary links, computed in reverse order along handle decompositions.","The skein module of B^4 recovers the original link homology, so the extension is a genuine enrichment of the input theory rather than a replacement.","Puncturing surfaces to make them framable decorates the resulting boundary links with canonical homology classes, yielding invariants of embedded and immersed surfaces and genus bounds.","The braided monoidal 2-category C_H plays the role for 4-manifold skein theory that ribbon categories play for 3-manifold skein theory, organizing the local relations of the theory.","Specific computations distinguish exotic smooth structures: the skein modules of the knot traces X_{-1}(-5_2) and X_{-1}(P(3,-3,8)) differ in quantum degree -1 over Q."],"supporting_citations":[{"why":"Main construction: Theorem 2.1 and the proof of the sweep-around move for gl_N homology.","marker":"[MWW22]"},{"why":"Introduces the gl_N link homology theories that the survey shows satisfy the extension hypotheses.","marker":"[KR08]"},{"why":"Establishes functoriality of colored link homologies in R^3, giving the chain-level functor CKhR_N.","marker":"[ETW18]"},{"why":"Supplies the Kauffman trick that the sweep-around proof categorifies.","marker":"[Kau87]"},{"why":"Introduces skein lasagna modules and the 2-handle formula, with computations for CP^2.","marker":"[MN22]"},{"why":"Develops reverse handle-decomposition computations and formulas for 1-handles and 3-handles.","marker":"[MWW23]"},{"why":"Uses skein modules to distinguish an exotic pair of knot traces and proves vanishing results.","marker":"[RW24]"},{"why":"Gives invariants of embedded and immersed surfaces and genus bounds from equivariant gl_N homology.","marker":"[MWW24]"},{"why":"Computes the skein lasagna module of S^2×S^2 and describes S^2×D^2, providing key test cases.","marker":"[SZ24]"}],"fun_headline_variants":["Link homology detects exotic smooth 4-manifolds","Skein lasagna modules distinguish exotic 4-manifolds","Link homology now tells smooth 4-manifolds apart","The 4-manifold invariant hidden in link homology","Link homology reaches into 4-manifold topology"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole tower of extensions collapses if the link homology fails to send some sweep-around move to the identity map; this is an infinite family of non-local moves, and the proof for gl_N homology is a delicate comparison of Reidemeister-3 chain maps.","fun_headline_variants_meta":{"raw":{"variants":["Link homology detects exotic smooth 4-manifolds","Skein lasagna modules distinguish exotic 4-manifolds","Link homology now tells smooth 4-manifolds apart","The 4-manifold invariant hidden in link homology","Link homology reaches into 4-manifold topology"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001576,"raw_usage":{"total_tokens":6274,"prompt_tokens":917,"completion_tokens":5357,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":533,"completion_tokens_details":{"reasoning_tokens":5276}},"tokens_in":533,"tokens_out":5357,"duration_ms":35866,"temperature":1.0,"reasoning_tokens":5276,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T16:08:30.320556+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a concrete tangle T in the sweep-around move—for example a single crossing or a braid closure—and compute the chain map that Khovanov homology (gl_2) assigns to the two movies related by that move; if the two maps are not homotopic, Theorem 2.4 and the whole extension fail. Alternatively, find a link cobordism in $S^{3}$×I whose movie passes through the point at infinity and whose induced map on KhR_N is not the identity.","supporting_citations":[],"review_version":2}