Quantum Master Equation and Open Gromov-Witten Theory 2

Vito Iacovino

REVIEW 3 major objections 5 minor 25 references

Quantum Master Equation and Open Gromov-Witten Theory 2

T0 review · 3 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A not-abelian open Gromov-Witten potential satisfies the quantum master equation up to master isotopy.

desk verdict The non-abelian OGW potential is a genuine new construction, but this paper is a conditional sketch leaning on the author's unpublished [5] and on unproved Proposition 30; still worth a specialist referee. read the letter →

arxiv 2412.06450 v1 pith:TKBCPJ5U submitted 2024-12-09 math.SG

Vito Iacovino This is my paper · ORCID

classification math.SG MSC 53D4553D37

keywords openGromov-Wittentheoryquantummasterequationmulti-curvechaincomplexdecoratedgraphsCalabi-YaumanifoldLagrangiansubmanifoldpartitionfunctionChern-Simonspropagator

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper defines a not-abelian open Gromov-Witten potential, a formal generating series that packages open Gromov-Witten invariants of a Calabi-Yau six-manifold with a Maslov-zero Lagrangian submanifold. The central claim is that the full partition function is the exponential of this potential, and that the potential satisfies the quantum master equation up to master isotopy. This gives a consistency condition for the formal invariants: the equation is what guarantees that the potential and the partition function are well defined under the auxiliary choices made in their construction. The construction rests on earlier results associating multi-curve cycles to moduli spaces of pseudo-holomorphic multi-curves and on a factorization property saying that disjoint unions of curves correspond to products of cycles.

What carries the argument

The load-bearing object is the multi-curve chain complex of decorated graphs. A graph has components labeled by a genus and a degree, and at each vertex the half-edges carry a cyclic order; the differential $\hat\partial = \partial + \delta + \eth$ encodes edge contraction, degenerate vertices, and moving marked points. The open Gromov-Witten cycles $Z_\beta$ are classes in this complex obtained from parametrized moduli spaces of multi-curves, following the cited construction in [2]. On top of this, the non-abelian propagator $P^{\mathrm{not-ab}} = P \otimes \mathrm{Id}$ with values in the Lie algebra, together with the trace over the cyclic order at each vertex, defines the forms $\Omega_{G,m}$; pairing $\Omega$ with cycles produces $P(Z)$ and $W(Z)$, and the BV operator $\Delta$ encodes the degenerations of marked points. The factorization property makes the product $\boxtimes$ compatible with $\hat\partial$ and forces $P = \exp(W/g_s)$.

What would settle it

Take two classes $\beta_1, \beta_2 \in H_2(X,L)$ and compute both sides of $\mathrm{fact}_{\beta_1,\beta_2}([Z_{\beta_1+\beta_2}])$ and $[Z_{\beta_1}] \boxtimes [Z_{\beta_2}]$ in the multi-curve chain complex of a concrete pair $(X,L)$; any nonzero difference would disprove the factorization property and with it the exponential formula. Alternatively, evaluate $W$ at the lowest orders in $g_s$ for a Lagrangian with known holomorphic-disk counts; a violation of $dW + \tfrac12\{W,W\} + g_s \Delta W = 0$ at the first nonconstant term would show the potentials defined here are not the physical open Gromov-Witten invariants.

Watch

Extended reading notes

Core claim

The author claims to associate to each class $\beta \in H_2(X,L)$ a not-abelian Gromov-Witten multi-curve cycle $Z_\beta$, built from moduli spaces of pseudo-holomorphic multi-curves, and then to integrate a trace of the Chern-Simons propagator over these cycles to get numbers $P(Z_\beta)$. Summing over $\beta$ with a Novikov variable gives the partition function $P(Z) = \sum_\beta P(Z_\beta) T^{\omega(\beta)}$, while restricting the sum to connected graphs gives the potential $W(Z)$. The paper's core statement is that the factorization property $\mathrm{fact}_{\beta_1,\beta_2}([Z_{\beta_1+\beta_2}]) = [Z_{\beta_1}] \boxtimes [Z_{\beta_2}]$ turns the partition function into an exponential, $P(Z) = \exp(W/g_s)$, and that isotopies of the cycles act on $W$ by master isotopies, moving solutions of the quantum master equation $dW + \tfrac12\{W,W\} + g_s \Delta W = 0$ to other solutions.

Load-bearing premise

The whole construction rests on the previously cited claims that the moduli spaces of holomorphic multi-curves give closed cycles in the graph complex and that these cycles multiply when the curves are disjoint; if either cited claim fails, the potential, the partition function, and the exponential relation are not defined.

Editorial extensions

If this is right

If the construction is correct, open Gromov-Witten invariants of $(X,L)$ organize into a single formal potential $W$, so invariants at all genera and degrees are constrained by the quantum master equation rather than being independent counts.
The exponential formula $P(Z) = \exp(W/g_s)$ means every disconnected contribution is determined by connected multi-curve invariants.
Invariance up to master isotopy means the potential is not a single function but a solution curve in the space of formal series; observable quantities must be built from objects invariant under the flow $dW + \tfrac12\{W,W\} + g_s \Delta W = 0$.
Bulk deformations do not require new degrees of freedom: adding a bulk parameter $A$ to the four-chain $K$ is equivalent to shifting the formal variable $b$, namely $P(\beta, K+rA, A)(g_s,b) = P(\beta, K, A)(g_s, b + r g_s)$.
The quantum master equation gives the consistency condition for the potential to be interpreted as the effective action of an open string field theory around the Lagrangian $L$.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The same structure may extend to other gauge groups and flat connections; the $U(N)$ trace used here is the simplest case, and the skein-style relations cited in the paper suggest a knot-invariant interpretation of the potential.
If the cited factorization property failed in some example, the exponential formula would still have a chance to hold after modifying the definition of connected invariants, since the partition function is the primary object and $W$ is derived from it.
A direct low-order computation of $W$ for a toric Calabi-Yau with a known Lagrangian, such as the resolved conifold with a Lagrangian $S^3$, would provide a concrete test of the master equation and of the factorization property.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Two things to know. First, the non-abelian OGW potential is a genuine new object: cyclic-ordered half-edge graphs, the non-abelian propagator P⊗Id, the trace over cyclic products, and the claim that the partition function is exp(W/gs). Second, the paper as posted is not self-contained enough to verify the central claim. You would need [5] (unpublished), Proposition 30 from [4], and Theorem 3 from [2] to check that W exists and satisfies QME. The current text reads more like a detailed research announcement than a proof.

What is good: the construction of the non-abelian MC chain complex is explicit, the algebra around the trace in (43) and the propagator in (45)–(46) is concrete, and the relation to point-splitting perturbative Chern–Simons in Section 2.7 is a useful bridge. This is not a routine “add colors” generalization; the cyclic order matters for the trace and for the skein relations.

Soft spots, in rough order of severity. The factorization property (Prop 30) is the load-bearing premise for P = exp(W/gs). It is cited, not proved. The stress-test note is right: this is not a formal consequence of the Leibniz rule; it is a geometric statement about disconnected moduli spaces, and if it fails, W does not exist. Relatedly, the coherent-cycle results from [5] (Props 18/21, Lemmas 22–25) are essential and unpublished. A referee cannot check this paper without that material. Lemmas 12 and 14 are stated without proof and are needed for the transfer map and for independence of isotopy. Prop 28/29 proofs are sketched in a few lines; they may be standard, but as written they are not complete. Finally, the partial order defining the induction in Section 2.6 has a self-comparison error: two bullets compare β(G′) to itself rather than to β(G). That is an easy typo to fix, but in the current version it undermines the printed induction.

Overall, the central idea is credible and the conditional chain is transparent. The paper is not circular in a bad sense, but the self-citation burden is heavy and one unpublished companion paper is essential. If the author posts [5] and fleshes out Prop 30, this would become a solid paper. For now it is a serious draft.

My recommendation: send it to a specialist referee, with explicit instructions that access to [5] and [4] is required. I would not desk reject it. But I would not cite it in my own work until the dependencies are public.

Referee Report

3 major / 5 minor

Summary. The paper defines a 'not abelian' multi-curve chain complex for a compact oriented 3-manifold with a finite-rank charge group, in which vertices carry cyclic orders of half-edges. It then defines an open Gromov-Witten partition function P(Z) by integrating a non-abelian Chern-Simons-like propagator over MC-cycles, and defines the open Gromov-Witten potential W(Z) as the connected-graph part. The main advertised results are that P satisfies dP + g_s ΔP = 0, that P(Z) = exp(W/g_s) when the cycles satisfy a factorization property, and hence that W satisfies the quantum master equation dW + (1/2){W,W} + g_s ΔW = 0. The paper is a companion to the author's previous and companion papers; the geometric input (existence of the non-abelian MC-cycles, factorization, and coherent cycles) is imported from [2], [4], and the unpublished [5].

Significance. If the imported geometric results hold, the paper would give an explicit chain-level construction of a non-abelian open Gromov-Witten potential satisfying the quantum master equation, with a transparent combinatorial mechanism at the level of decorated graphs. The trace formula (43)–(47) and the statement of Lemma 31 are clear, and the paper correctly identifies factorization as the key structural input needed to pass from a partition function to a connected potential. The main advertised result, however, is conditional on external results that are not proved or made available in this manuscript; this makes the current submission, as it stands, incomplete as an independent proof of the central claim.

major comments (3)

[§3.2, Prop. 30] The exponential formula P((Z_β)_β) = exp(W((Z_β)_β)/g_s) at the start of §3.2, and therefore the quantum master equation for W at the end of §3.2, rest entirely on Proposition 30, which asserts the factorization property fact_{β1,β2}([Z_{β1+β2}]) = [Z_{β1}] ⊠ [Z_{β2}]. This proposition is not proved in the present paper; it is only cited to [4]. The factorization identity is not a formal consequence of the chain-level Leibniz rule for ⊠: it is a geometric statement about the class of the disconnected moduli-space cycle in the non-abelian graph complex, where vertices carry cyclic orders and the weight (43) is defined through cyclic traces. Unless Proposition 30 is proved here or its proof in [4] is reproduced in a verifiable form, the main identified consequence for W is not established by this manuscript.
[§2.7.1, Props. 18, 21 and Lemmas 22–25] These results are all cited to '[5]' (in preparation). They provide the coherent MC-cycles Z(w,fr,U▲), the normalized MC-chain complex C‡, the skein map from Skein(M)[[g_s,a]]_+ to MCH(M)‡, and the universal power series A, β, α, r, θ. Even if the skein relations are not directly used in the final QME for W, the coherent cycles and the normalization C‡ are presented as part of the framework of nice MC-cycles used throughout §2.6–§2.7, and Proposition 18–21 are advertised constructions of the paper. As written, this substantial part of the paper is unverifiable without access to an unpublished manuscript, and the dependence should either be eliminated or the results should be included with full proofs.
[§3, Props. 28 and 29] The proof of Proposition 28 is only a sketch. The key identities dΩ_{G,m} = ∑_{m' | ∂_{l+1}m' = m} Ω_{G,m'} and ΔΩ_{G,m} = ∑_{m' | ∂_0 m' = m} Ω_{G,m'} are stated without derivation, and the sign bookkeeping involving the orientation o(H(G)) and the anti-invariance (42) of the propagator is not shown. Proposition 29 is dismissed with 'can be proved as the last Proposition.' Since these propositions are the mechanism that turns the chain-level master equation into the PDE for P and then for W, the main technical step should be written out completely or given a precise, verifiable reference.

minor comments (5)

[§2.6] The definition of the partial order on decorated graphs contains a typographical error: in the first two bullets, ω(β(G')) appears on both sides of the inequality and equality, so the intended comparison with ω(β(G)) is lost; this should be corrected.
[§3, Eq. (47)] The finiteness of the sum defining P(Z) is not discussed. Since the chain complex C_*(β) in (4) is not equipped with an explicit finite-support condition, the paper should specify in which formal power series ring the infinite sum over decorated graphs converges, or add a support condition to the definition of a MC-chain.
[§3] The BV operator Δ and the odd bracket { , } are used in equations (47)–(48) but are never defined in this paper. For a self-contained presentation, a precise definition or an exact reference to [1] or [4] should be given.
[Throughout] There are numerous typographical errors, including 'simplectic' in the introduction, 'Gromow-Witten' in §3.2, and 'reacher' for 'richer' in the introduction; the manuscript would benefit from a careful proofreading pass.
[§2.2.2] The phrase 'Given two MC-cycles Z0 and Z0' should read 'Z0 and Z1'; the notation is otherwise clear but this typo is confusing.

Circularity Check

2 steps flagged · score 4.0 of 10

The not-abelian potential and its QME are conditionally derived from the author's prior Proposition 30 and Theorem 3; the paper's own calculations are not equation-level circular.

self citation load bearing [Section 3.1 (Proposition 30) and Section 3.2 (exponential formula)]
"Proposition 30. ([4]) To the moduli space of multicurves we can associate a collection of nice multi-curve cycles (Zβ )β with Zβ ∈ Z β which satisﬁes the factorization property. ... The factorization property and Lemma 31 imply that the open Gromov-Witten partition function is the exponential of the open Gromov-Witten potential P(Z) = exp( 1 gs W(Z))."

The central identity for the not-abelian potential, P(Z) = exp(W/gs), is not proved in this paper from the chain complex and propagator; it is made to follow by invoking Proposition 30, stated verbatim as a result of the author's own earlier paper [4] and given no proof here. The factorization property is precisely the multiplicative input needed to reorganize the graph sum into connected components. If that property fails, the exponential formula and the resulting QME dW + 1/2{W,W} + gs ΔW = 0 do not follow. Thus this load-bearing step reduces to a self-citation rather than to a derivation contained in the present paper.
self citation load bearing [Section 2.5, Theorem 3]
"Theorem 3. ([2]) Let β ∈ H2(X, L, Z). To the moduli space of pseudoholomotphic multi-curves of homology class β it is associated a multi-curve cycle Zβ ∈ Z β |UK of Euler class [UK]."

The partition function P(Z) and the potential W(Z) are defined by pairing the forms ΩG,m with the components of an MC-cycle Z. The existence of the non-abelian MC-cycles Zβ is the geometric input on which the entire construction rests. The paper gives no proof of this existence or of the asserted isotopy-independence; it quotes Theorem 3 from the author's own prior paper [2]. All subsequent QME statements are conditional on this imported, same-author result.

full rationale

Within the paper itself, the algebraic derivation is not circular: Proposition 28 proves P(∂hat B) = gs Δ P(B), Proposition 29 proves the QME for P, Lemma 31 proves multiplicativity of P under the product Z1 ⊠ Z2, and the passage from P to the connected-graph potential W is a standard exponential/log reorganization once factorization is granted. No parameter is fitted and no prediction is renamed as an input. However, the geometric existence and factorization of the not-abelian MC-cycles—Theorem 3 from [2] and Proposition 30 from [4]—are load-bearing and are imported from the author's own prior papers without proof here. Because these citations are not machine-checked or replaced by an in-paper argument, they constitute partial self-citation load-bearing, though [2] and [4] are publicly posted and give the paper independent content beyond a bare renaming. The additional results cited to the unpublished [5] concern coherent cycles and skein relations; these support the PSPCS comparison more than the central QME claim. Score 4, not higher, because the QME reduction and the integration pairing are original to this paper and would be valid consequences if the cited geometric inputs hold.

Assumptions & free parameters 0 free parameters · 4 assumptions · 2 invented entities

The central claim rests on the author's prior results for MC-cycles and factorization, plus unpublished data for coherent cycles and skein series. There is no empirical fitting; formal variables are bookkeeping. The invented entities are mathematical definitions whose external validation is deferred to companion papers.

assumptions (4)

domain assumption Theorem 3 ([2]): To the moduli space of pseudoholomorphic multi-curves of class β it is associated a multi-curve cycle Zβ ∈ Zβ|UK of Euler class [UK], well-defined up to isotopy.
Central input to the whole construction; proved in the author's previous paper [2], not in the present text.
domain assumption Proposition 30 ([4]): The collection of nice MC-cycles (Zβ)β satisfies the factorization property factβ1,β2([Zβ1+β2]) = [Zβ1] ⊠ [Zβ2].
Quoted from companion paper [4]; used to derive P = exp(W/gs).
ad hoc to paper Existence of coherent cycles Z(w,fr,U▲) and skein power series A, β, α, r, θ (Propositions 18/21, Lemmas 22-25), cited to [5], in preparation.
The reference [5] is not publicly available; these results are assumed for the Chern-Simons/skein interpretation.
ad hoc to paper The partial order on decorated graphs is well-founded (Section 2.6).
The definition as printed compares ω(β(G′)) with itself, so the intended area/Euler-characteristic filtration is not actually defined; the induction used in Lemmas 4-8 depends on this order.

invented entities (2)

Non-abelian Multi-Curve chain complex C†, C‡
purpose: Foundational algebraic structure used to define the non-abelian open Gromov-Witten potential
Introduced in Section 2; its relation to actual moduli spaces is quoted from the author's earlier papers, not demonstrated here.
Coherent M C-cycles Z(w,fr,U▲)
purpose: Canonical generators of the normalized MC-homology used for skein relations
Defined in [5] (in preparation), not available; no independent confirmation.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Quantum Master Equation and Open Gromov-Witten Theory 2." pith.science (2026). https://pith.science/paper/TKBCPJ5U

@misc{pith2026241206450,
  author       = {Pith},
  title        = {Pith review of: Quantum Master Equation and Open Gromov-Witten Theory 2},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TKBCPJ5U}},
  note         = {Machine review of arXiv:2412.06450}
}

read the original abstract

We define the not abelian Open Gromov-Witten potential.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

25 extracted references · 12 canonical work pages

[5]

Gaiotto, G

D. Gaiotto, G. W. Moore and A. Neitzke, Four-dimensional wall-crossing via threedimensional field theory, arXiv:0807.4723

arXiv
[4]

Cordova, S

C. Cordova, S. Espahbodi, B. Haghighat, A. Rastogi, and C. Vafa, Tangles, generalized Reidemeister moves, and thre-dimensional mirror symmetry, arXiv:1211.3730

arXiv
[2]

Cecotti and C

S. Cecotti and C. Vafa, BPS Wall Crossing and Topological Strings , arXiv: 0910.2615

arXiv
[1]

Bullimore, T

M. Bullimore, T. Dimofte, D. Gaiotto, and J. Hilburn, Boundaries, Mirror Symmetry, and Symplectic Duality in 3d N = 4 Gauge Theory, arXiv:1603.0838

work page arXiv
[3]

Cecotti, C

S. Cecotti, C. Cordova, and C. Vafa, Braids, Walls, and Mirrors , arXiv:1110.2115

arXiv
[6]

Gaiotto, G

D. Gaiotto, G. W. Moore and A. Neitzke, Wall-crossing, Hitchin Systems, and the WKB Approximation, arXiv:0907.3987

arXiv
[7]

Gaiotto, G

D. Gaiotto, G. Moore, A. Neitzke, Framed BPS States , arXiv:1006.0146

arXiv
[8]

Gaiotto, G

D. Gaiotto, G. Moore, A. Neitzke, Wall-Crossing in Coupled 2d-4d Systems, arXiv:1103.2598

arXiv

Show all 25 references

[9]

Fukaya; K

K. Fukaya; K. Ono, Arnold conjecture and Gromov-Witten invariant, Topology 38 (1999), no. 5, 933--1048

1999
[10]

Fukaya, Y.-G

K. Fukaya, Y.-G. OH, H. Ohta, K. Ono, Lagrangian intersection Floer theory - anomaly and obstruction
[11]

Kontsevich and Y

M. Kontsevich and Y. Soibelman, Stability Structures, Motivic Donaldson-Thomas Invariants and Cluster Transformations , arXiv:0811.2435

arXiv
[12]

Iacovino, Master Equation and Perturbative Chern-Simons theory, arXiv:0811.2181

V. Iacovino, Master Equation and Perturbative Chern-Simons theory, arXiv:0811.2181

arXiv
[13]

Iacovino, Open Gromov-Witten theory on Calabi-Yau three-fold I, arXiv:0907.5225

V. Iacovino, Open Gromov-Witten theory on Calabi-Yau three-fold I, arXiv:0907.5225

arXiv
[14]

Iacovino, Open Gromov-Witten theory on Calabi-Yau three-folds II, arXiv:0908.0393

V. Iacovino, Open Gromov-Witten theory on Calabi-Yau three-folds II, arXiv:0908.0393

arXiv
[15]

Iacovino, Frame ambiguity in Open Gromov-Witten invariants, arXiv:1003.4684

V. Iacovino, Frame ambiguity in Open Gromov-Witten invariants, arXiv:1003.4684

arXiv
[17]

Iacovino, Kontsevich-Soibelman Wall Crossing Formula and Holomorphic Disks, arXiv:1711.05306

V. Iacovino, Kontsevich-Soibelman Wall Crossing Formula and Holomorphic Disks, arXiv:1711.05306

arXiv
[18]

Iacovino, Open Gromov-Witten theory without Obstruction, arXiv:1711.05302

V. Iacovino, Open Gromov-Witten theory without Obstruction, arXiv:1711.05302

arXiv
[19]

Iacovino, Open Gromov-Witten invariants and Bounrdary States, arXiv:1807.08786

V. Iacovino, Open Gromov-Witten invariants and Bounrdary States, arXiv:1807.08786

arXiv
[20]

Iacovino, Quantum Master Equation and Open Gromov-Witten theory, arXiv:2412.04230

V. Iacovino, Quantum Master Equation and Open Gromov-Witten theory, arXiv:2412.04230

arXiv
[21]

Iacovino, Point Splicting Perturbative Chern Simons, in preparation

V. Iacovino, Point Splicting Perturbative Chern Simons, in preparation
[22]

Iacovino, Holomorphic Lagrangian Floer Homology, in preparation

V. Iacovino, Holomorphic Lagrangian Floer Homology, in preparation
[23]

Iacovino, Refined Wall Crossing and Open Holomorphic Curves, in preparation

V. Iacovino, Refined Wall Crossing and Open Holomorphic Curves, in preparation
[24]

Ooguri, C

H. Ooguri, C. Vafa, Knot invariants and topological strings, Nuclear Phys. B 577 (2000) 419–438 MR1765411

2000
[25]

Witten, Chern-Simons Gauge Theory As A String Theory, arXiv:hep-th/9207094

E. Witten, Chern-Simons Gauge Theory As A String Theory, arXiv:hep-th/9207094

arXiv
[26]

Zaslow, Wavefunctions for a Class of Branes in Three-space, arXiv:1803.02462

E. Zaslow, Wavefunctions for a Class of Branes in Three-space, arXiv:1803.02462

arXiv

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.

[1] [5]

Gaiotto, G

D. Gaiotto, G. W. Moore and A. Neitzke, Four-dimensional wall-crossing via threedimensional field theory, arXiv:0807.4723

arXiv

[2] [4]

Cordova, S

C. Cordova, S. Espahbodi, B. Haghighat, A. Rastogi, and C. Vafa, Tangles, generalized Reidemeister moves, and thre-dimensional mirror symmetry, arXiv:1211.3730

arXiv

[3] [2]

Cecotti and C

S. Cecotti and C. Vafa, BPS Wall Crossing and Topological Strings , arXiv: 0910.2615

arXiv

[4] [1]

Bullimore, T

M. Bullimore, T. Dimofte, D. Gaiotto, and J. Hilburn, Boundaries, Mirror Symmetry, and Symplectic Duality in 3d N = 4 Gauge Theory, arXiv:1603.0838

work page arXiv

[5] [3]

Cecotti, C

S. Cecotti, C. Cordova, and C. Vafa, Braids, Walls, and Mirrors , arXiv:1110.2115

arXiv

[6] [6]

Gaiotto, G

D. Gaiotto, G. W. Moore and A. Neitzke, Wall-crossing, Hitchin Systems, and the WKB Approximation, arXiv:0907.3987

arXiv

[7] [7]

Gaiotto, G

D. Gaiotto, G. Moore, A. Neitzke, Framed BPS States , arXiv:1006.0146

arXiv

[8] [8]

Gaiotto, G

D. Gaiotto, G. Moore, A. Neitzke, Wall-Crossing in Coupled 2d-4d Systems, arXiv:1103.2598

arXiv