{"id":"1129f46d-61fb-4549-aafd-356f313a4a90","arxiv_id":"2505.14889","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"An explicit inverse matrix A_{u,beta} is constructed for braid varieties; its frozen columns generate the cluster automorphism group action, and the matrix is proven invertible with determinant plus or minus one.","lead":"This paper describes the automorphism group of the cluster algebra attached to a braid variety, giving an explicit matrix that controls how the automorphisms scale the cluster variables. The matrix is shown to be invertible with integer entries, and the action is worked out in examples, including a newly observed sign pattern.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 3.7's Case A induction hinges on Lemma 3.15's exclusion of boundary patterns (0,1,0) and (1,0,1), asserted only by a citation to [17]; if those patterns can occur, the reflection step and the determinant theorem fail.","rationale":"The reader's weakest-assumption analysis and my independent reading converge on Lemma 3.15: the unproved exclusion of the two boundary patterns is the point where the determinant theorem's induction is least secure. The matrix construction and the Case B row/column reduction are otherwise plausible, and the worked examples are consistent with det \\tilde B = ±1. The orientation mismatch between Definition 2.4/Remark 2.25 and Lemma 4.2/Corollary 4.3 is an internal convention issue: the Corollary uses the transpose convention appropriate for Lam–Speyer's Proposition 5.1, so it is fixable without changing the mathematical conclusion. Lemma 3.18's omitted direct computation is a rigor gap but less fundamental than the unverified local exclusion, since the latter underpins the entire Case A induction. If the proposed enumeration confirms the exclusion, I would regard the determinant theorem as sound modulo the requested exposition fixes; if it produces a counterexample, the main theorem is false. Therefore I keep the reader's CONDITIONAL verdict rather than moving to ACCEPT or REJECT, and I recommend the concrete check above as the decisive test.","tokens_in":19012,"tokens_out":19991,"duration_ms":176613,"concrete_test":"Enumerate all small pairs (u,β) with n ≤ 5 and ℓ(β) ≤ 6 for which Case A applies, using Galashin's code or an independent implementation of [17]. For every soap film C_j of (u',β'), check the triple (∂_{u',β'}(C_j)_{i−1}, ∂_{u',β'}(C_j)_i, ∂_{u',β'}(C_j)_{i+1}) at the distinguished index i. If (0,1,0) or (1,0,1) occurs, Lemma 3.15 is false and Theorem 3.7 needs a different Case A argument. Alternatively, locate the exact statement in [17] or the positive distinguished subexpression literature and verify that it applies to the opposite quiver and vertex ordering used in this paper.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central determinant theorem is proved by induction, and the Case A step (β = σ_i β', u = s_i u') reduces \\tilde B_{u,β} to \\tilde B_{u',β'} through Lemma 3.15, which asserts ∂_{u,β}(C_j) = R_i(∂_{u',β'}(C_j)). The proof of Lemma 3.15 excludes exactly the length-three patterns (0,1,0) and (1,0,1) at positions (i−1,i,i+1), citing only the 'positive distinguished subexpression' property from [17] and giving no derivation or precise pointer. This exclusion is load-bearing: if such a pattern occurred in ∂_{u',β'}, then R_i would force an entry 2 or −1, contradicting that boundary maps take values in {0,1}. That would invalidate the equality D_{u,β} = D_{u',β'} and hence the Case A determinant reduction. It is not a general fact about soap films: Example 3.10 shows isolated 1s in boundary maps, so the exclusion must depend on the specific distinguished subexpression. The paper does not establish this property, and it is not obviously implied by the material in §2.3. Separately, Lemma 3.18's 'direct computation' is not written out, and a sign error there would change equation (3) and the determinant sign in Case B. Since Corollary 4.3 and the torus action depend on Theorem 3.7 and really full rank, a failure of Lemma 3.15 would also invalidate the main application. The paper's orientation inconsistency between Definition 2.4 and Lemma 4.2 is real but appears to be a fixable convention issue; it does not reduce the weight of the unproved local exclusion.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper defines an (m+f) x (m+f) integer matrix \\tilde B_{u,beta} for each braid variety X_{u,beta}, obtained by adding a boundary correction matrix D_{u,beta} to the half-arrow matrix H_{u,beta} of the 3D plabic graph. The main theorem (Theorem 3.7) states that this matrix has integer entries and determinant (-1)^{m+f}, and the paper proves it by induction over the braid word, splitting into a Case A where the first letter of beta removes a simple reflection from u and a Case B where it does not. From the determinant statement and Lam-Speyer's Proposition 5.1, the paper derives that the cluster automorphism group Aut(A(Q_{u,beta})) is an algebraic torus (C*)^f and gives an explicit action on the initial seed x_i in Corollary 4.3. The last section computes several examples, exhibits a 'sign phenomenon' for the inverse matrix A_{u,beta}, and gives a counterexample to that phenomenon.","tokens_in":19311,"tokens_out":7200,"duration_ms":61964,"significance":"If the proof is completed, the paper provides a concrete, computable description of the cluster automorphism group for braid varieties, complementing the general framework of Lam-Speyer and giving a practical way to compute the torus action on cluster variables. The matrix construction is well adapted to 3D plabic graphs and the examples are computed in detail, including verification by Sage in one case and use of Galashin's program in others. The inductive factorization of the inverse matrix A_{u,beta} in Lemma 4.5 is a useful structural addition. The main limitation is that two load-bearing local statements in the proof of Theorem 3.7 are not fully demonstrated in the text, so the determinant theorem and its corollaries are not yet fully self-contained.","major_comments":[{"comment":"The proof of Lemma 3.15 excludes exactly the triples (1,0,1) and (0,1,0) in the boundary vector on the grounds that they 'do not happen due to the fact that we choose a positive distinguished subexpression for u inside beta, see [17]' (page 14). This exclusion is load-bearing for Case A of Theorem 3.7: without it, applying R_i to a boundary vector with a 1 at position i would produce entries 2 or -1, contradicting that boundary maps take values in {0,1} and invalidating the equality D_{u,beta}=D_{u',beta'} used in the induction. The property is not a general fact about all 3D plabic graphs, since Example 3.10 contains the boundary vector (0,1,0). The authors should either prove the forbidden-pattern statement from the positive distinguished subexpression property within the paper, or quote and verify the precise statement in [17] that implies it. As written, the determinant theorem is not fully self-contained.","section":"Section 3.2, Lemma 3.15"},{"comment":"The displayed identity H'_{y',x'} = H_{y,x} + H_{m+1,x}D_{m+1,y} - D_{m+1,x}H_{m+1,y} is justified only as a 'direct computation' from Figure 5, with the intermediate algebra involving the multiplicities a_x,b_x,c_x,d_x omitted. This identity is used to derive equation (3) in the proof of Theorem 3.7 and hence controls the determinant sign in Case B; a sign error in this computation would change the final determinant value (-1)^{m+f}. The authors should include the coordinate calculation, or at least a table of the seven cases with the corresponding values of the four multiplicities, so that the identity can be verified by the reader.","section":"Section 3.2, Lemma 3.18"},{"comment":"There is a mismatch in the shape of the exchange matrix that affects the proof of the main application. Definition 2.4 defines \\tilde B(Q) as an (m+f) x m matrix, while the proof of Lemma 4.2 and the multiplication map in Corollary 4.3 require \\tilde B(Q) to be an m x (m+f) matrix after identifying its rows with the mutable part of \\tilde B_{u,beta}. As written, the expression \\tilde B(Q)v for v in Z^{m+f} is not defined under Definition 2.4. This is fixable by a transposition or by explicitly defining \\tilde B(Q) to be the first m rows of \\tilde B_{u,beta}, but it should be corrected because Lemma 4.2 is the bridge between the determinant theorem and the description of the cluster automorphism group.","section":"Definition 2.4 and Lemma 4.2"}],"minor_comments":[{"comment":"In Remark 2.16, 'C_{m+1}, ..., C_{m+f} corresponds to a mutable vertex' should read 'frozen vertex'; this typo is confusing in a passage that is supposed to fix the vertex labeling.","section":"Remark 2.16"},{"comment":"In Case 2 of Lemma 3.19, the string 'a2+1' should be 'a_{i+2}', and the boundary vectors should be displayed with consistent indexing; the same typo appears in the surrounding cases.","section":"Lemma 3.19, Case 2"},{"comment":"In the displayed action for Example 4.7, the expression for x1 contains the factor t_3^{-1} twice; the matrix row for x1 indicates the intended action includes each of t_1^{-1}, t_2^{-1}, t_3^{-1}, t_4^{-1}, t_5^{-1} exactly once.","section":"Example 4.7"},{"comment":"The Introduction relies on the unpublished manuscripts [1] and [4] for the statement that the extended exchange matrices from [2] and [17] are the same. The authors should clarify whether any of the paper's main claims depend on these forthcoming works, and if so, state the precise assertions that are being used.","section":"Introduction"}],"recommendation":"major_revision","confidential_remarks":"The main determinant theorem is plausible and the overall strategy is sensible, but the current version contains two proof gaps in load-bearing lemmas (Lemma 3.15 and Lemma 3.18) and a matrix-convention inconsistency that affects the statement and proof of Lemma 4.2 and Corollary 4.3. These are fixable within the manuscript's scope, so I recommend major revision rather than rejection. I would also ask the editor to ensure the authors specify how much of the argument depends on the in-preparation references [1] and [4], since the text currently cites them for a non-negligible claim about cluster structures."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The short version: the paper proves a genuinely new determinant theorem for the extended exchange matrix \\tilde{B}_{u,\\beta} attached to a braid variety, and uses it to make the cluster automorphism group and its action explicit. Theorem 3.7 is not in the existing literature, and the inductive proof is mostly careful. The inverse factorization in Lemma 4.5 is a real computational handle, and the worked examples are detailed enough to verify the statements, including a counterexample to the 'sign phenomenon' the author observes. That part is solid.\n\nThe soft spots are not fatal, but they are real. Lemma 3.15 is load-bearing: the Case A induction reduces the boundary matrix by a reflection, and the equality of boundary corrections depends on excluding the length-three patterns (0,1,0) and (1,0,1) in the boundary map. The proof says this follows from the positive distinguished subexpression property in [17], but gives no derivation and no precise pointer. If those patterns can occur, the reflection step would force entries 2 or -1 and the determinant theorem would fail. I do not see that this is obviously implied by the §2.3 material; it needs a proof or a quotable statement from [17]. Similarly, Lemma 3.18's 'direct computation' from Figure 5 is not shown; it is a plausible identity, but a sign error there would change the determinant in Case B. Both are referee-fixable, but they are exactly where a careful referee should spend time.\n\nThe orientation inconsistency between Definition 2.4 and Lemma 4.2/Corollary 4.3 is real but cosmetic: once the transposed exchange matrix convention is reconciled, the application goes through. The reliance on unpublished works [1,4] for the equivalence of cluster structures is worth flagging, though the determinant theorem concerns the 3D plabic graph construction of [17], so it is not circular.\n\nWho this is for: researchers working on cluster structures on braid varieties or on automorphism groups of cluster algebras. It is not a broad-audience breakthrough, but it is a credible, checkable contribution that advances the concrete computational side of the program. It deserves a serious referee. I would send it out and ask for the Lemma 3.15 proof and the Lemma 3.18 computation to be written out.\n\nRecommendation: accept with major revisions, conditional on those gaps being closed.","headline":"Useful new determinant theorem and explicit automorphism group action for braid variety cluster structures, but the main proof leans on one under-justified combinatorial exclusion that a referee should demand be closed.","tokens_in":19904,"tokens_out":2993,"would_cite":true,"duration_ms":25138,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["13F60","14M15","20F36"],"pacs":[],"model":"deepseek-v4-flash","headline":"For every braid variety, the cluster automorphism group is an algebraic torus whose action on the cluster variables is given explicitly by the inverse of an extended exchange matrix with determinant $\\pm 1$.","keywords":["cluster algebra","cluster variety","braid variety","cluster automorphism group","3D plabic graph","soap film","extended exchange matrix","algebraic torus"],"falsifier":"Compute $\\partial(C_j)$ for every soap film in the 3D plabic graph of a braid pair $(u,\\beta)$: finding a film whose boundary vector contains a consecutive triple $(0,1,0)$ or $(1,0,1)$ at the crossing added in Case A would refute Lemma 3.15 and with it the induction in Theorem 3.7. Alternatively, compute $\\det \\tilde{B}_{u,\\beta}$ for any positive braid word and check that it is not $(-1)^{m+f}$; the theorem predicts this value in all cases.","tokens_in":18723,"feed_emoji":"🧶","tokens_out":11330,"duration_ms":91557,"temperature":0.7,"pith_summary":"The paper gives an explicit description of the cluster automorphism group of a braid variety $X_{u,\\beta}$, the affine variety attached to a permutation $u$ and a positive braid word $\\beta$. The main technical claim is that the extended exchange matrix $\\tilde{B}_{u,\\beta}$ obtained by adding a boundary correction term to the half-arrow matrix of a 3D plabic graph has integer entries and determinant $(-1)^{m+f}$. Since this makes $\\tilde{B}_{u,\\beta}$ invertible over the integers, the inverse matrix $A_{u,\\beta}$ becomes the key object: its last $f$ columns form a basis of the kernel of the exchange matrix, and each basis vector supplies the exponents of one $\\mathbb{C}^\\times$ factor of the automorphism group. The paper's corollary is that $\\operatorname{Aut}(\\mathcal{A}(Q_{u,\\beta}))$ is an algebraic torus of dimension $f$ acting on the braid variety by the monomial scalings $x_i \\mapsto \\prod_j t_j^{a_{i,m+j}} x_i$. This matters because it converts a question about symmetries of cluster algebras into linear algebra on one explicitly computed integer matrix, and the paper's examples show the torus can be larger than the familiar $(n-1)$-dimensional action from flag geometry.","feed_headline":"Braid-variety symmetries are explicit tori","feed_subtitle":"An extended exchange matrix with determinant ±1 yields the full automorphism group and its action on cluster variables.","key_machinery":"The load-bearing object is the extended exchange matrix $\\tilde{B}_{u,\\beta} = H_{u,\\beta} + D_{u,\\beta}$. Here $H_{u,\\beta}$ is the half-arrow matrix of the 3D plabic graph of $(u,\\beta)$---the planar projection of a spatial graph whose bridges and regions encode the cluster seed---recording arrows between quiver vertices with half-integer weights when both endpoints are frozen. The boundary correction matrix $D_{u,\\beta}$ has entries $(\\partial(C_i), \\partial(C_j))$ measuring how the soap films $C_i, C_j$ (the regions attached to the bridges) cover the $n-1$ boundary regions, paired by a symmetric bilinear form built from the negative half Cartan matrix of type $A_{n-1}$. The proof of Theorem 3.7 shows by induction on the length of $\\beta$ that adding $D_{u,\\beta}$ cancels every half-integer and forces the determinant to be exactly $(-1)^{m+f}$. The inverse matrix $A_{u,\\beta}$ then carries the application: its last $f$ columns are the kernel basis whose entries become exponents in the torus action.","core_discovery":"On the paper's own terms, the discovery is that for every braid variety $X_{u,\\beta}$ the cluster automorphism group $\\operatorname{Aut}(\\mathcal{A}(Q_{u,\\beta}))$---the group of algebra automorphisms that send every cluster variable to a nonzero scalar multiple of itself---is an algebraic torus $(\\mathbb{C}^\\times)^f$, where $f$ is the number of frozen vertices of the quiver built from the 3D plabic graph of $(u,\\beta)$. The action is explicit: if $A_{u,\\beta} = \\tilde{B}_{u,\\beta}^{-1}$ and $\\operatorname{col}_j(A_{u,\\beta}) = (a_{1,m+j}, \\ldots, a_{m+f,m+j})$ for $m+1 \\le j \\le m+f$, then the torus acts on an initial seed by $x_i \\mapsto \\prod_{j=1}^f t_j^{a_{i,m+j}}\\, x_i$. The determinant identity $\\det \\tilde{B}_{u,\\beta} = (-1)^{m+f}$ is the load-bearing fact: it guarantees that the last $f$ columns of the inverse form an integer basis of the kernel of the exchange matrix, which Proposition 5.1 of [13] identifies with the automorphism group.","pith_inferences":["A testable extension is to scan the same matrix construction over all braid words up to a fixed length and map exactly where the sign phenomenon fails; the paper's examples suggest it may persist for double Bott–Samelson-type quivers but not for quivers with mixed arrows.","Because the argument only uses the really-full-rank property of the quiver and the graph combinatorics, the inverse-matrix recipe for the automorphism group should transfer to any ice quiver arising from a 3D plabic graph, not necessarily from a braid variety.","The explicit monomial action makes the fixed loci of the torus computable, which could be used to test the cluster deep locus and no mysterious point predictions on braid varieties mentioned in the introduction."],"forward_implications":["The cluster automorphism group of every braid variety is an algebraic torus of dimension equal to the number of frozen vertices of the 3D plabic quiver.","The torus action on the braid variety is monomial, with exponents read off from the last $f$ columns of $A_{u,\\beta}$, so it can be computed directly from the graph.","The inductive factorization of $A_{u,\\beta}$ in Lemma 4.5 gives a row-and-column-operation recipe for the automorphism group without passing through the whole mutation class.","For $u = \\mathrm{id}$ the torus is $(\\mathbb{C}^\\times)^l$ and coincides with the known action, while for general $u$ the dimension $f$ can exceed $n-1$, so the automorphism group is genuinely larger than the standard flag action.","The observed sign phenomenon---nonzero entries of $A_{u,\\beta}$ all sharing one sign---holds in the running examples but fails in Example 4.8, so it is not a general theorem; the paper poses the problem of describing the nonzero entries combinatorially."],"supporting_citations":[{"why":"Supplies the 3D plabic graph and soap film construction, the cluster structure theorem for braid varieties, and the positive distinguished subexpression fact used in Lemma 3.15.","marker":"[17]"},{"why":"Supplies the definition of cluster automorphism and Proposition 5.1 identifying the automorphism group with the kernel of the multiplication map, which is the bridge to the torus action.","marker":"[13]"},{"why":"Provides the alternative cluster structure on braid varieties and the matrix $(p_{i,j})$ that motivates $\\tilde{B}_{u,\\beta}$; Lemma 8.1 there connects $\\tilde{B}_{u,\\beta}$ to the cluster Poisson structure.","marker":"[2]"},{"why":"Defines braid varieties and provides the $(\\mathbb{C}^\\times)^{n-1}$ action that the paper compares and contrasts with the full automorphism group.","marker":"[3]"},{"why":"Introduced the cluster automorphism group, the object whose explicit form the paper computes for braid varieties.","marker":"[18]"}],"fun_headline_variants":["Braid-variety automorphism group is an explicit torus","Automorphism group of braid varieties: a torus","Torus action on braid varieties made explicit","Braid variety automorphisms: explicit tori","Cluster automorphisms of braid varieties are tori"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof assumes that when a crossing is added to the braid word, a soap film never has the boundary pattern $(0,1,0)$ or $(1,0,1)$ at that spot; the paper relies on a cited fact for this, and if such a pattern occurred the reflection step could produce boundary entries $2$ or $-1$, breaking the induction.","fun_headline_variants_meta":{"raw":{"variants":["Braid-variety automorphism group is an explicit torus","Automorphism group of braid varieties: a torus","Torus action on braid varieties made explicit","Braid variety automorphisms: explicit tori","Cluster automorphisms of braid varieties are tori"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.0006,"raw_usage":{"total_tokens":2767,"prompt_tokens":870,"completion_tokens":1897,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":486,"completion_tokens_details":{"reasoning_tokens":1817}},"tokens_in":486,"tokens_out":1897,"duration_ms":13159,"temperature":1.0,"reasoning_tokens":1817,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T15:28:16.428433+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute $\\partial(C_j)$ for every soap film in the 3D plabic graph of a braid pair $(u,\\beta)$: finding a film whose boundary vector contains a consecutive triple $(0,1,0)$ or $(1,0,1)$ at the crossing added in Case A would refute Lemma 3.15 and with it the induction in Theorem 3.7. Alternatively, compute $\\det \\tilde{B}_{u,\\beta}$ for any positive braid word and check that it is not $(-1)^{m+f}$; the theorem predicts this value in all cases.","supporting_citations":[{"cited_title":"Cohomology of cluster varieties, I: Locally acyclic case","cited_arxiv_id":null,"evidence_quote":"Supplies the definition of cluster automorphism and Proposition 5.1 identifying the automorphism group with the kernel of the multiplication map, which is the bridge to the torus action."},{"cited_title":"Cluster structures on braid varieties","cited_arxiv_id":null,"evidence_quote":"Provides the alternative cluster structure on braid varieties and the matrix $(p_{i,j})$ that motivates $\\tilde{B}_{u,\\beta}$; Lemma 8.1 there connects $\\tilde{B}_{u,\\beta}$ to the cluster Poisson structure."},{"cited_title":"Algebraic weaves and braid varieties","cited_arxiv_id":null,"evidence_quote":"Defines braid varieties and provides the $(\\mathbb{C}^\\times)^{n-1}$ action that the paper compares and contrasts with the full automorphism group."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduced the cluster automorphism group, the object whose explicit form the paper computes for braid varieties."}],"review_version":1}