{"id":"d0c4286e-e91c-4f23-b381-5a8c0186ff74","arxiv_id":"2506.09972","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"The paper derives a tree-of-unlinkings formula for wild Donaldson-Thomas invariants of m-Kronecker quivers from wall-crossing identities rewritten through symmetric quivers and diagonalization.","lead":"This paper rewrites Kontsevich-Soibelman wall-crossing identities for m-Kronecker BPS quivers, including wild ones, as equalities between generating series of symmetric quivers that represent dual 3d N=2 boundary theories. This yields a formula expressing closed Donaldson-Thomas invariants of the 4d theories in terms of open invariants and m-loop quiver invariants, with low-order checks.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Equation (6.57) lacks a proof that its tree enumeration is complete and multiplicity-free; the (c - r^T) factors are an ordered falling-product count whose binomial normalization is explicitly left open, and the low-order checks never probe branching with repeated identical endpoints.","rationale":"The reader's weakest assumption is that the tree enumeration is complete and that (6.57) uniquely fixes the closed invariants. I agree this is the load-bearing point, and I isolate a concrete unresolved mechanism: the (c - r^T) factors are falling products for ordered selections, while the text defines trees up to permutation of equivalent endpoint nodes. That distinction cannot be tested by the Appendix B examples, which only use single-endpoint trees or c=1 multiplicities. This supports keeping the CONDITIONAL verdict: the low-order matches are real evidence, but the general formula is not yet established. I do not propose rejection, because the construction is coherent, the low-order checks pass, and the missing piece is a combinatorial normalization and an explicit triangularity argument rather than a demonstrated contradiction. A targeted computation at (2,2) for m=3 would settle whether the overcount is real or whether the falling product is actually the correct ordered-diagonalization count.","tokens_in":63686,"tokens_out":6811,"duration_ms":97528,"concrete_test":"Implement an exhaustive backtracking enumeration of unlinking trees for m=3 (then m=4) up to open charge (B,C)=(2,2), applying rules (3.12)-(3.13) and the tree definitions of Section 6.2. Extract the coefficient of x_2^2 x_1^2 q^{D/2} from (6.57) under two hypotheses: (A) the falling-product factors (c - r^T_{pq}) as written; (B) binomial factors choosing r^T_{pq} nodes from c identical endpoints. Compare both to the coefficient obtained directly from the equality P_{Qs}=P_{Qw} truncated at that charge, and to independent wild DT data for m=3 from Reineke or attractor-flow computations. The variant that matches identifies the missing symmetry factor; if neither matches, the tree enumeration itself is incomplete.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that (6.57) determines the wild closed DT invariants requires the right-hand side to be an exact, multiplicity-free enumeration of the diagonalization of Q_w, and the resulting equations to be triangular with unique solutions. Neither is demonstrated. The concrete issue is the factor (Omega^{4d}_k(gamma)_{ip} - r^T_{pq}) in (6.57), inherited from the product in (6.42). For a tree that uses the same equivalence class of c_{ip,dip,kip} identical nodes d_p times, that product is c(c-1)...(c-d_p+1), an ordered falling factorial. But the paper defines trees up to permutation of sequences with the same endpoint multiplicities ('several sequences ... can be combined into the same sequence and defined as the same tree T'), so the correct factor would be binomial unless the sum over T explicitly orders the endpoints. The paper itself flags this as open: after (6.44) it says it would be interesting to see whether the product can be arranged into binomial coefficients. The appendix checks only n=1 trees or c=1 nodes, so those checks cannot detect an overcount. In addition, since Q_w's adjacency matrix contains the unknown c's, the recursive definitions of S and T in Section 6.2 are self-referential: completeness of the tree set and the asserted order-by-order solvability of (6.57) are not proven.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a dictionary between both sides of Reineke's motivic wall-crossing identities for m-Kronecker quivers and symmetric quivers of 3d N=2 boundary theories. In the strong-coupling chamber the symmetric quiver is the doubled m-Kronecker quiver; in the weak-coupling chamber the authors construct an infinite symmetric quiver Q_w whose adjacency matrix is expressed in terms of the unknown closed Donaldson-Thomas invariants c^{a,b}_k. The central result is equation (6.57), which claims that 3d open BPS degeneracies of the doubled quiver equal a sum over trees of unlinkings of products (Omega^{4d}_k(gamma)_{i_p} - r^T_{pq}) times m-loop quiver invariants. The paper uses low-order cases for m=3,4,6 to recover known Reineke values for rays of the form (1,k).","tokens_in":64082,"tokens_out":5635,"duration_ms":73027,"significance":"If correct, the construction gives a new open-closed relation between 4d class S wall-crossing and 3d vortex partition functions, and a potentially practical route to computing wild Donaldson-Thomas invariants. The explicit normal-ordering derivation of the Q_w adjacency matrix and the reproduction of the q-binomial values for m=3,4,6 are concrete strengths of the paper. However, the central combinatorial step, namely the tree enumeration leading to (6.42)-(6.57), is not yet a well-defined, complete, and multiplicity-free enumeration, and the claimed order-by-order solvability is asserted rather than proved. The paper is best viewed as a promising proposal whose central claim needs substantial rigor before it can be accepted.","major_comments":[{"comment":"The completeness of the tree enumeration is declared rather than proven: no argument shows that every sequence of unlinkings producing a given final identification appears exactly once in S, nor that the infinite sums over T, g_T, and l_{\\mu\\nu} converge or truncate at each order. Because the adjacency matrix of Q_w contains the unknown invariants c^{a,b}_k, the membership of a tree in S depends on the very quantities the equations are meant to determine. In particular, the claim that (6.56) and (6.57) can be solved order by order requires a triangularity and uniqueness proof that is not supplied.","section":"§6.2, sets S, T, ST; equations (6.15)-(6.22)"},{"comment":"The factor (c^{...}-r^T_{pq}) counts ordered choices, while several endpoints of a tree are allowed to be in the same equivalence class as in (6.35). For d_p identical endpoints, the product in (6.42) is therefore an ordered falling factorial c(c-1)...(c-d_p+1), whereas the tree is defined up to permutation of equivalent endpoints; the correct multiplicity-free count would be binomial unless the sum over T explicitly orders the endpoints. The paper explicitly leaves this issue open in the sentence after (6.44), and the checks in Appendices B.2-B.4 do not exercise a branching tree with repeated identical endpoints, so they cannot detect an overcount. Equation (6.57) is thus not yet a well-defined enumeration.","section":"§6.2, equations (6.42)-(6.44)"},{"comment":"The integers r^T_{pq} are said to depend on how many times the tree returns to a given equivalence class and on which specific nodes have previously been unlinked, but no algorithm is given to compute them for an arbitrary tree. Without such a prescription, the product in (6.42) and the condition t^T_{pq}>0 in (6.43) are not fully specified, and different assignments of r^T_{pq} could change the solution set for c^{a,b}_k.","section":"§6.2, definition of r^T_{pq}"},{"comment":"The paper asserts that comparing open DT invariants yields equations that 'can be effectively solved', but it does not prove that the infinite system has a unique solution. The appendix examples merely reproduce known invariants for m=3,4,6; they demonstrate consistency with known results, not uniqueness of the solution to (6.57). If that system admits multiple solutions satisfying all low-order checks, the central claim that it determines wild DT invariants would fail.","section":"§6.3, equation (6.57)"}],"minor_comments":[{"comment":"There are several typos in the introduction, including 'fined means', 'Lagangian', and 'characerizes'; the manuscript would benefit from a careful proofreading pass.","section":"§1"},{"comment":"The adjacency matrices of Q_w, especially in §4.2 and §5, are rendered in a way that is very hard to read; the authors should consider a cleaner typesetting or an ancillary file with the full matrices.","section":"§4.2 and §5"},{"comment":"The symbol C is used both for the adjacency matrix and for the dense cone of BPS rays; this overloaded notation is confusing in places such as §4.2 and §6.1 and should be disambiguated.","section":"Throughout"},{"comment":"The sign conventions and powers in the identifications (4.15)-(4.19) are stated without derivation; a brief indication of how they follow from the normal-ordering computation in Appendix A would improve readability.","section":"§4.2, equations (4.15)-(4.19)"},{"comment":"Reference [49] is listed as 'To appear' with no further information; since the paper relies on it for geometric interpretation of the symmetric quiver map, the authors should provide an arXiv number or a more complete citation.","section":"§7"}],"recommendation":"major_revision","confidential_remarks":"The paper fits the journal's scope and contains a genuinely interesting idea, but the central combinatorial claim is not yet rigorous. In my view the main obstacle is not the physical setup but the definition and normalization of the tree sum in (6.57) and the proof that the resulting equations determine c^{a,b}_k uniquely. I would ask the authors to provide a precise combinatorial prescription for ordered trees or binomial counts, and to test the formula on at least one branching configuration with repeated identical endpoints before resubmission. The reliance on the authors' earlier diagonalization paper [59] is legitimate and does not itself raise concerns."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Cut to the chase: this is a serious paper with a good new idea, but the central formula (6.57) is not yet proven in the regime it's designed for. The low-order checks are real and reassuring; the gaps are mostly combinatorial, not physical.\n\nWhat's new: the reformulation of Reineke's wild quantum dilogarithm identity as an equality of symmetric quiver generating series, the construction of the infinite symmetric quiver Q_w, and the tree-of-unlinkings formula that expresses closed 4d DT invariants in terms of open 3d invariants and m-loop quiver invariants. That's a genuinely new angle, and the appendix normal-ordering computations are substantial. The m=3,4,6 examples reproduce Reineke's q-binomial coefficients, which gives solid evidence the machinery works at low order.\n\nThe soft spots are real but not fatal. The most concrete issue is the combinatorial normalization in (6.57). The product over q of (c - r^T_{pq}) is an ordered falling factorial, while the trees are defined up to permutation of identical endpoints. Unless the tree sum explicitly orders the endpoints, that's an overcount. The authors flag this themselves right after (6.44) and leave it open. The appendix checks never probe it—they only use n=1 trees or c=1 nodes, so a missing factorial would go unnoticed. Second, the completeness of the tree enumeration is asserted, not proven. The recursive definitions of S and T in Section 6.2 are self-referential, and there's no argument that every contribution to an open invariant is captured exactly once. That's load-bearing. Third, the order-by-order solvability of the equations is plausible but unproven; no triangularity or uniqueness argument is offered.\n\nNone of this makes the paper nonsense. It's honest about the first gap, and the low-order checks are a legitimate checkpoint. But the central formula isn't yet established in the wild regime. For peer review: send it, but demand a rigorous treatment of the tree enumeration and the factorial/binomial normalization, or at least an explicit check at an order where the issue would surface.\n\nThis is for people working on BPS quivers, DT theory, and wall-crossing. It deserves a serious referee.","headline":"Promising reformulation of wild wall-crossing via symmetric quivers; low-order checks pass, but the central tree formula has an unproven counting factor that the authors themselves leave open.","tokens_in":64582,"tokens_out":5760,"would_cite":false,"duration_ms":64456,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["16G20","14N35","81T60"],"pacs":[],"model":"deepseek-v4-flash","headline":"Wild BPS degeneracies in 4d N=2 theories can be computed from a tree-sum formula built from 3d symmetric quiver data.","keywords":["wall-crossing","BPS quivers","symmetric quivers","DT invariants","wild quivers","m-Kronecker quivers","quantum dilogarithm identities","3d-4d duality"],"falsifier":"Compute a higher-order wild invariant, for instance the m = 3 quiver at charge (3,1), by an independent method such as spectral networks or the split attractor flow formula and compare it with the value obtained by solving the tree equations order by order; a mismatch would show the tree enumeration is incomplete.","tokens_in":63507,"feed_emoji":"⚛️","tokens_out":6069,"duration_ms":66997,"temperature":0.7,"pith_summary":"This paper claims that wild wall-crossing, the regime where BPS states condense into a dense cone of rays, can be tamed by translating 4d BPS quiver identities into equalities of symmetric quivers for 3d boundary theories. On one side of the wall sits the symmetrized BPS quiver; on the other sits an infinite symmetric quiver whose nodes correspond to the weak-coupling chamber's BPS states. By diagonalizing both quivers, the paper derives a formula expressing each 3d open BPS number as a sum over trees of unlinkings of products of 4d closed wild BPS numbers and m-loop quiver invariants. Solving these equations determines unknown closed invariants order by order, which the paper demonstrates for low-order examples with m = 3, 4, and 6 arrows.","feed_headline":"Tree sums tame wild BPS wall-crossing","feed_subtitle":"Diagonalizing symmetric quivers turns unknown 4d degeneracies into solvable equations, verified for m = 3, 4, 6.","key_machinery":"The machinery is the passage from a 4d BPS quiver to a symmetric quiver, meaning a quiver with equal arrow multiplicities in both directions, which encodes a 3d N=2 vortex theory through its generating series. Symmetrizing the m-Kronecker quiver gives the finite quiver $Q_s$, while normal-ordering the infinite product on the weak-coupling side of the wall yields an infinite symmetric quiver $Q_w$ whose adjacency matrix is determined up to the unknown closed invariants. Quiver diagonalization repeatedly applies the unlinking operation, which removes one pair of arrows between two nodes and creates a new node with loops while preserving the generating series, until all remaining nodes carry only loops. The tree formula counts all sequences of unlinkings that lead to the same final identification of variables, and equating the diagonalizations of $Q_s$ and $Q_w$ fixes the unknown coefficients.","core_discovery":"The central claim is equation (6.57): $$\n\\$\\Omega$^{3d}_{d,\\tilde{k}} = \\sum_{T_{d,\\tilde{k}}, $C^{{\\mathrm{loop}}$}_T} \\prod_{p,q} \\left(\\$\\Omega$^{4d}_{k}(\\gamma)_{i_p} - r^T_{pq}\\right) \\$\\Omega$^{3d}_{$C^{{\\mathrm{loop}}$}_T},\n$$ where the sum runs over trees of unlinkings that produce a given open BPS state, the factors track how many nodes of each charge remain ununlinked, and the endpoints carry invariants of m-loop quivers. The paper argues that since the left side is known from the symmetrized m-Kronecker quiver, this identity recursively fixes the unknown wild Donaldson-Thomas invariants $c^{a,b}_k$. The authors verify the mechanism on low-order coefficients for m = 3, 4, and 6, recovering known binomial values, and conjecture that the same open-closed relation extends to all 4d N=2 theories whose spectra are governed by a motivic wall-crossing formula.","pith_inferences":["If the tree enumeration is complete, the same diagonalization strategy should apply to any wall-crossing identity expressible as a product of quantum dilogarithms, not just m-Kronecker quivers, giving a general open-closed relation for BPS degeneracies.","The tree-sum structure resembles split attractor flow trees, so matching the two tree formalisms could yield a dictionary between unlinking combinatorics and supergravity flow trees, making closed invariants computable by either method.","The low-order examples suggest the recursion may admit a closed-form expression for $c^{a,b}_k$ in terms of binomial coefficients; finding such a formula would extend the known results for charges of the form $(1,b)$ to the whole dense cone."],"forward_implications":["Wild Donaldson-Thomas invariants for m-Kronecker quivers with m >= 3 can be computed recursively, order by order, from the tree formula rather than from spectral networks or attractor-flow trees.","Wall-crossing identities for BPS quivers become dualities of 3d N=2 theories: the symmetrized BPS quiver and the infinite quiver $Q_w$ have identical vortex partition functions after variable identification.","Products of quantum dilogarithms in wall-crossing formulas reduce to combinatorial identities among symmetric quiver generating series, giving a 3d interpretation of the 4d spectrum jump.","Low-order computations for m = 3, 4, and 6 reproduce known q-binomial results, indicating the recursive scheme is consistent and extendable to higher charges and other spins."],"supporting_citations":[{"why":"Supplies the wild quantum dilogarithm identity with unknown coefficients $c^{a,b}_k$ that the paper reformulates as symmetric quiver identities.","marker":"[50]"},{"why":"Defines symmetric quiver generating series and the linking and unlinking operations from multicover skein relations used throughout the paper.","marker":"[12]"},{"why":"Introduces quiver diagonalization, the algorithm the paper applies to both sides of the wall to extract Donaldson-Thomas invariants.","marker":"[59]"},{"why":"Provides the original motivic wall-crossing framework in terms of quantum dilogarithms that the paper recasts into 3d language.","marker":"[2]"},{"why":"Gives the Donaldson-Thomas invariants of m-loop quivers, which appear as the endpoint contributions in the tree formula.","marker":"[44]"},{"why":"Presents the split attractor flow tree approach that the paper expects to match its trees of unlinkings.","marker":"[40]"}],"fun_headline_variants":["Symmetric quivers crack wild BPS degeneracies","Wild 4d BPS from 3d open invariants","Tree unlinkings turn wild quivers tractable","Wall-crossing via diagonalized m-Kronecker","Open-closed formula for wild DT invariants"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that no sequence of unlinkings is missing from the tree sum and that the resulting equations have a unique solution for the unknown BPS numbers.","fun_headline_variants_meta":{"raw":{"variants":["Symmetric quivers crack wild BPS degeneracies","Wild 4d BPS from 3d open invariants","Tree unlinkings turn wild quivers tractable","Wall-crossing via diagonalized m-Kronecker","Open-closed formula for wild DT invariants"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000194,"raw_usage":{"total_tokens":1371,"prompt_tokens":980,"completion_tokens":391,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":596,"completion_tokens_details":{"reasoning_tokens":327}},"tokens_in":596,"tokens_out":391,"duration_ms":5025,"temperature":1.0,"reasoning_tokens":327,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T04:36:49.201441+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute a higher-order wild invariant, for instance the m = 3 quiver at charge (3,1), by an independent method such as spectral networks or the split attractor flow formula and compare it with the value obtained by solving the tree equations order by order; a mismatch would show the tree enumeration is incomplete.","supporting_citations":[{"cited_title":"Wild quantum dilogarithm identities","cited_arxiv_id":"2302.12062","evidence_quote":"Supplies the wild quantum dilogarithm identity with unknown coefficients $c^{a,b}_k$ that the paper reformulates as symmetric quiver identities."},{"cited_title":"Quiver diagonalization and open BPS states","cited_arxiv_id":"2212.04379","evidence_quote":"Introduces quiver diagonalization, the algorithm the paper applies to both sides of the wall to extract Donaldson-Thomas invariants."},{"cited_title":"Degenerate Cohomological Hall algebra and quantized Donaldson-Thomas invariants for m-loop quivers","cited_arxiv_id":"1102.3978","evidence_quote":"Gives the Donaldson-Thomas invariants of m-loop quivers, which appear as the endpoint contributions in the tree formula."}],"review_version":1}