{"id":"d45b2eb1-451d-4954-a5e4-3d914622e49f","arxiv_id":"2506.03252","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Exponential and logarithmic integrals are shown to be twisted periods, with explicit wall-crossing and thimble decompositions for a Pearcey integral and an elliptic family, proposed as the geometric origin of master-integral decompositions in Feynman integrals.","lead":"This paper recasts exponential and logarithmic integrals, including Feynman integrals in the Baikov representation, as periods of twisted cohomology with explicit thimble decompositions. It derives wall-crossing data for a Pearcey integral and an elliptic-curve example, and proposes that thimble bases match decompositions into master integrals.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Equation (4.74) is internally inconsistent: with (4.72)–(4.73), the long exact sequence forces dim H^2(pair)=4−r and dim H^3(pair)=2−r, so the printed (C^2,C^4) cannot occur.","rationale":"The reader's weakest assumption focused on the omitted kernel of d^1 in Section 4.3; that is a fair reproducibility concern. My independent check of the same section shows a stronger, more direct problem: even granting (4.72), the final LES result is arithmetically impossible. This is not an appeal to consensus or an ad hominem; it is a finite linear algebra check from equations printed in the paper. The first line of (4.74) checks out, so the Re γ>0 direction is fine, and the master-integral/decomposition proposal may still be salvageable in a companion paper. But the concrete example's headline two-dimensionality for Re γ<0 is not established, and the stated C^4 in H^3 is inconsistent. Hence the current version should be rejected or heavily revised; if the authors supply the rank computation and correct (4.74), a conditional acceptance could be reconsidered.","tokens_in":45124,"tokens_out":18270,"duration_ms":187309,"concrete_test":"Recompute (4.74) by exact dimension bookkeeping: take the LES (4.32) with (4.72) and (4.73) and determine the rank r of the connecting map H^2(X,L)→H^2(D_R^B,L). Concretely, compute both groups as group cohomology of π_1(D_R^B) (the Heisenberg group) for the representation ρ, write the restriction map on 2-cocycles, and evaluate its rank; then report dim H^2(pair)=4−r and dim H^3(pair)=2−r. This single check decides whether the corrected second line is 0⊕0⊕C^4⊕C^2⊕0 (r=0), 0⊕0⊕C^3⊕C^1⊕0 (r=1), or 0⊕0⊕C^2⊕0⊕0 (r=2); no value of r yields 0⊕0⊕C^2⊕C^4⊕0.","verdict_should_be":"REJECT","load_bearing_attack":"The central claim that the elliptic example has two-dimensional middle cohomology for both signs of Re γ rests on (4.74). For Re γ<0 the second line of (4.74) reads H^•(~X,D_R^B,Π_*L)=0⊕0⊕C^2⊕C^4⊕0. But substituting the paper's own (4.72) H^•(D_R^B,L)=0⊕C^2⊕C^2⊕0 and (4.73) H^•(X,L)=0⊕0⊕C^2⊕0⊕0 into the long exact sequence (4.32) gives, with r=rank(H^2(X,L)→H^2(D_R^B,L)), dim H^2(pair)=4−r and dim H^3(pair)=2−r at H^2 and H^3. Since r∈{0,1,2}, the possible pairs are (4,2), (3,1), (2,0); (2,4) is impossible. Thus the printed result cannot follow from the stated inputs, independent of the omitted d^1 kernel flagged by the reader. If r=0 the middle H^2 is C^4, not C^2; if r=2 then H^3=0, not C^4. The first line of (4.74) is consistent. The two-dimensionality claim for Re γ<0 is therefore currently unsupported.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper applies the Kontsevich–Soibelman framework for exponential integrals to several physics applications. After reviewing twisted de Rham and Betti cohomology for triples (X,D0,f), it studies the Pearcey integral relevant to gauged Skyrme models, computing thimbles, vanishing cycles, intersection numbers, and Stokes jump matrices for the three discriminant regimes. It then extends the formalism to closed 1-forms α = d log B, proposes this as a framework for Baikov Feynman integrals, and works out the example of the Legendre elliptic family B = y^2 + x(x−1)(x−λ), computing the global Betti cohomology via the long exact sequence of the pair and identifying thimble periods with elliptic integrals. The main claims are that the thimble decomposition aligns with the decomposition into master integrals and that the wall-crossing structure gives a sharp count of independent master integrals.","tokens_in":45378,"tokens_out":11080,"duration_ms":115554,"significance":"The ambition is high and the payoff would be significant if the claims hold: a geometric, parameter-free organization of master-integral decompositions and of their analytic continuation in dimensional-regularization parameters. The paper contains several concrete and checkable computations: the Pearcey monodromy matrices (3.54) and (3.63), the jump matrices (3.58) and (3.66), and the elliptic monodromies (4.58) with intersection form (4.59). These are not fitted and are not circular, and the review of [1] is useful for physicists. However, the central elliptic computation contains an inconsistency that invalidates the stated two-dimensionality result for Re γ < 0, and other load-bearing steps are asserted rather than proved. At present the main claims are therefore not established.","major_comments":[{"comment":"The second line of (4.74) is inconsistent with Eqs. (4.72), (4.73) and the long exact sequence (4.32). Write A^k = H^k(X,L), B^k = H^k(DR_B, Π_*L), and P^k = H^k(∼X, DR_B, Π_*L). From (4.73), A^0 = A^1 = A^3 = A^4 = 0 and A^2 = C^2; from (4.72), B^0 = B^3 = 0 and B^1 = B^2 = C^2. The LES segment B^1 → P^2 → A^2 → B^2 → P^3 → A^3, together with P^1 = 0, forces B^1 → P^2 to be injective. Writing r = rank(A^2 → B^2), one obtains dim P^2 = 4 − r and dim P^3 = 2 − r. With r ∈ {0,1,2}, the possible pairs are (4,2), (3,1), and (2,0); the printed pair (2,4) cannot occur. Thus (4.74) cannot follow from the stated inputs. The two-dimensionality of the middle cohomology for Re γ < 0 is therefore unsupported: if r = 2 then H^3 vanishes rather than being C^4, and if r < 2 then H^2 has dimension at least three. This is a load-bearing error for the paper's central claim.","section":"§4.3, Eq. (4.74)"},{"comment":"The computation of H•(DR_B, Π_*L) is not actually shown: the kernel of d^1 in the Heisenberg group cohomology is stated to be 'tedious but direct' and left to the reader. This is a load-bearing step, since (4.72) is the boundary input B^• in the long exact sequence that determines the final cohomology in (4.74). The full computation of ker d^1 and the image of d^0 must be supplied, or a precise reference to a lemma in [1] must be given. In addition, the reduction H^j(π_1(DR_B), V_{ρ*}) ≃ H^j(π_1(DR_B), V_ρ) for j = 0,1, stated after (4.66), needs justification; for group cohomology such an isomorphism is not automatic.","section":"§4.3, Eqs. (4.71)–(4.72)"},{"comment":"The replacement B^{−γ} → e^{−γ log B} for non-rational γ is not a trivial reformulation. The exponential e^{−γ log B} is multi-valued when γ is not an integer, and the statement that this 'solves the issue' with γ ∉ Q requires a precise definition of the branch and a proof that the Baikov integral equals the resulting exponential period. This is load-bearing for the claimed applicability to Feynman integrals in dimensional regularization, where γ is generically not rational. The same issue affects Eq. (4.21), where e^{−γ f_i} is only locally defined. The paper asserts this identification rather than proving it.","section":"§4, Eq. (4.2)"},{"comment":"The text acknowledges that the convergence condition for the thimble integral may fail when dim_C X ≥ 3. For Baikov representations of multi-loop Feynman integrals, the integration dimension is not bounded by two, so this is not a marginal restriction. The introductory claim that the strategy is not restricted by spacetime dimensions or special assumptions on the underlying geometry should be reconciled with this limitation, or the claim should be qualified.","section":"§4.2, after Eq. (4.22); §1"},{"comment":"The headline claim that the thimble decomposition aligns with the decomposition into Master Integrals is not demonstrated. In the elliptic example the authors compute cohomology dimensions and identify the resulting periods with elliptic integrals, but they do not perform any integration-by-parts reduction or exhibit a master-integral basis; the 'alignment' remains a proposed correspondence. Similarly, the claim that the wall-crossing structure allows a sharp count of independent Master Integrals is not substantiated in the example: the Stokes rays are given in (4.75), but the Stokes automorphisms and jump matrices for the elliptic family are not computed.","section":"Abstract, §4.3, §5"}],"minor_comments":[{"comment":"The right-hand side of (4.64) contains a stray period: the matrix should read [[2,1],[1,1]] rather than [[2,1],[1,1.]].","section":"§4.3, Eq. (4.64)"},{"comment":"The sentence 'Finally, we can you use (4.62),(4.72) and (4.73)' is ungrammatical and should read 'we can use'.","section":"§4.3, before Eq. (4.74)"},{"comment":"The abbreviation 't.i.' should be 'i.e.'.","section":"§2, after Eq. (2.7)"},{"comment":"The phrase 'let proceed extending' should be 'let us proceed by extending'.","section":"§3.4"},{"comment":"The caption of Figure 7 refers to 'Stokes’ lines' on the Cγ plane, while the text defines Stokes rays in Cγ and Stokes lines in Cn; please make the terminology consistent.","section":"Figure 7 caption and §3.2"},{"comment":"The coefficient c_{i,λ} in (2.40) is defined as a γ-independent expansion coefficient, but the expression in (2.50) has γ on the right-hand side; this should be clarified or corrected.","section":"§2.3, Eq. (2.50)"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is heavily programmatic and leans on [1] and on the companion paper [70] for essential details. In its current form, the central elliptic computation is internally inconsistent (see major comment 1), and the missing d^1 kernel computation is a serious gap. The authors should be asked to recompute Eq. (4.74) and to include the omitted group-cohomology details before the paper can be considered further. There is no indication of misconduct; the inconsistency looks like an indexing or rank error, but it is load-bearing."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: the Pearcey section is a real, self-contained computation worth having, but the elliptic Betti calculation in Section 4.3 does not add up. The stress-test note is right: feeding (4.72) and (4.73) into the long exact sequence (4.32) forces the pair cohomology for Re γ < 0 to have dims (4−r, 2−r) in degrees 2 and 3, where r is the rank of H^2(X,L) → H^2(D_R^B,L). That gives possible pairs (4,2), (3,1), or (2,0), never (2,4). So the printed second line of (4.74) cannot follow from the paper's own inputs, independent of the omitted d^1 kernel. The two-dimensionality claim for Re γ < 0 is unsupported as written. The first line may be fine, but the second line is the one carrying the 'independent of sign of γ' conclusion.\n\nWhat is genuinely new: the Pearcey jump matrices (3.58), the monodromy matrices S (3.60) and (3.67), and the explicit wall-crossing structure for the positive/negative discriminant cases. Those look carefully done and form a useful concrete application of the Kontsevich–Soibelman formalism to a physically motivated integral. The writing is clear, and the paper is honest that the Feynman-to-master-integral correspondence is a proposal deferred to a companion paper. The Baikov reformulation with log B is a reasonable framing, not a finished theorem, and it is labeled as such.\n\nSoft spots, in order: (1) the (4.74) inconsistency; (2) the kernel of d^1 in (4.71) is left to the reader, and the result (4.72) is stated without enough detail; (3) the identification of the Baikov exponential integral for non-rational γ is asserted rather than proved—fine as a proposal, but it is a leap; (4) the heavy reliance on [1] is acceptable since [1] is the mathematical foundation, but it means the paper does not stand alone as a proof of the master-integral program.\n\nBottom line: this is a serious paper that deserves refereeing, but the elliptic example needs a major repair. If the authors fix (4.74) and provide the missing kernel computation, the Pearcey part alone justifies publication. I would not cite the elliptic result in its current form.","headline":"The Pearcey wall-crossing computation is solid and new, but the elliptic Betti result has an internal inconsistency that undermines the central two-dimensionality claim.","tokens_in":45995,"tokens_out":5691,"would_cite":false,"duration_ms":51974,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["32S40","14D07","81T18"],"pacs":[],"model":"deepseek-v4-flash","headline":"Feynman integrals in Baikov form are exponential periods whose thimble decomposition matches master integrals, with wall crossings counting the independent ones.","keywords":["exponential integrals","twisted de Rham cohomology","Betti homology","wall-crossing structure","Baikov representation","Feynman integrals","master integrals","vanishing cycles"],"falsifier":"Compute the kernel of the differential $d^{1}$ in the group cohomology of the nilmanifold boundary for the elliptic example: if the twisted cohomology of that boundary is not 0 ⊕ $C^{2}$ ⊕ $C^{2}$ ⊕ 0, then the middle cohomology in equation (4.74) would not be $C^{2}$ and the match with two master integrals would fail. A more global test would be to apply the cohomological count to a known one-loop Baikov family, such as the two-point integral, and check that the number of independent periods equals the number of master integrals obtained by integration by parts.","tokens_in":44859,"feed_emoji":"📐","tokens_out":9837,"duration_ms":106113,"temperature":0.7,"pith_summary":"This paper tries to establish that a broad family of physical integrals, from saddle-point exponential integrals to Feynman integrals, can be understood as periods pairing twisted de Rham cohomology with Betti homology. For Feynman integrals, the multivalued logarithm of the Baikov polynomial plays the role of the exponent, making the integral an exponential period for any complex value of the dimensional-regularization parameter. The paper claims that the decomposition of these integrals over Lefschetz thimbles reproduces the standard master-integral decomposition, and that the wall-crossing structure in the auxiliary parameter plane gives a clean count of independent master integrals while avoiding Stokes ambiguities. The central worked example, an integral over a one-parameter family of elliptic curves, is computed explicitly: the middle cohomology is two-dimensional independently of the sign of the real part of the parameter, matching the two expected basis integrals. This matters because it would replace integration-by-parts reduction with a cohomological description of where master integrals come from.","feed_headline":"Wall-crossing geometry counts Feynman master integrals","feed_subtitle":"Baikov integrals recast as exponential periods; a two-dimensional middle cohomology matches two master integrals.","key_machinery":"The load-bearing object is the triple (X,D0,α), with X the complement of the zero locus of the Baikov polynomial, D0 the divisor controlling boundaries, and α the closed 1-form d log B. The covariant derivative ∇α = d − α∧ defines a twisted de Rham complex, while the Betti side is a local system of relative homology; the integral is the pairing of a class in each. Thimbles are the integration cycles: each starts at a zero of α and follows a level set of Im(γ log B), and their jumps across Stokes rays are encoded by Stokes automorphisms that define the wall-crossing structure. In the elliptic example the Betti computation also uses the real oriented blow-up of $P^{2}$ along the divisors, whose boundary over the elliptic curve is a circle bundle, a three-dimensional nilmanifold; the monodromy matrices of the local system on that boundary determine the twisted cohomology via group cohomology.","core_discovery":"The discovery is that the exponential-integral formalism, originally developed for a holomorphic exponent, extends to the multivalued setting needed for Feynman integrals by replacing the exact form df with a closed algebraic 1-form α. In the Baikov representation one takes α = d log B, where B is the Baikov polynomial, and the integration cycles are Lefschetz thimbles built from vanishing cycles of the fibers of B; the exponential period pairing is then defined for every γ in C*. The paper's main claim is that the thimble decomposition of the Baikov integral coincides with the master-integral decomposition, and that the Stokes automorphisms between adjacent sectors form a wall-crossing structure whose generic sector determines the number of independent master integrals. The elliptic example makes this concrete: the global Betti cohomology is computed from the monodromy of the elliptic family, and equation (4.74) gives middle cohomology $C^{2}$ for both Re γ > 0 and Re γ < 0, so the period space has exactly two dimensions, matching the two master integrals.","pith_inferences":["This suggests a topological shortcut for master-integral counting in multi-scale families: compute the rank of the local system over a generic sector rather than running integration-by-parts reduction, with jumps across walls read off from Stokes automorphisms.","The same cohomological translation should apply to other parametric representations, such as Schwinger or Feynman parameters, wherever the integrand is a power of a polynomial; one test is whether the resulting master-integral counts agree with known integration-by-parts counts for one-loop N-point integrals.","The convergence condition noted for complex dimension at least three implies that not every thimble pairing is automatically a master integral; this predicts that for high-dimensional Baikov examples some basis elements require additional boundary data, a claim that could be checked numerically.","If the wall-crossing description extends to families with kinematic parameters, it gives a geometric picture of discontinuities in master-integral bases across thresholds, potentially unifying threshold expansions with the Stokes sectors."],"forward_implications":["Baikov Feynman integrals become exponential periods for every complex γ, so dimensional regularization, including irrational γ, no longer forces one to leave the geometric period picture.","The number of independent master integrals equals the dimension of the relevant twisted cohomology in a generic sector, giving a principled count that does not depend on choosing contours near Stokes lines.","A basis of thimbles provides explicit integration cycles whose integrals are the master integrals; at large γ their expansion coefficients are periods of the ordinary cohomology of the algebraic fibers.","In the elliptic example, the two-dimensional middle cohomology matches exactly two basis integrals, supporting the identification of thimble decomposition with master-integral decomposition."],"supporting_citations":[{"why":"Supplies the twisted de Rham and Betti framework for exponential integrals, including the local-to-global isomorphisms, Stokes rays, and wall-crossing structure that the paper adapts.","marker":"[1]"},{"why":"Establishes the identification of Feynman integrals with periods and intersection theory, the background against which master-integral decomposition is recast cohomologically.","marker":"[29]"},{"why":"Gives the Baikov representation of Feynman integrals, the starting point for writing the integrand as e^{-γ log B}.","marker":"[90]"},{"why":"Provides the review of the Baikov representation and integration by parts that motivates counting master integrals.","marker":"[91]"},{"why":"Supplies the physical quartic-exponential integral from a nuclear-matter partition function used as the first worked example.","marker":"[65]"},{"why":"Introduces the wall-crossing structure in supersymmetric theories that the exponential-integral Stokes automorphisms generalize.","marker":"[84]"},{"why":"Used to classify the singular fiber types of the elliptic example, fixing the local monodromy data needed for the Betti cohomology computation.","marker":"[92, 93]"}],"fun_headline_variants":["Wall-crossing reveals number of Feynman master integrals","Exponential periods align Feynman integrals with wall-crossing","Thimble decomposition matches Feynman master integral expansion","Wall-crossing structure counts independent Feynman periods"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing assumption is that the formal global-to-local isomorphisms between twisted de Rham and Betti cohomologies hold for the Baikov pair (X,D0,d log B); in the concrete elliptic example, the claimed two-dimensional middle cohomology also depends on a group-cohomology kernel computation for the nilmanifold boundary that the authors state is tedious but direct and leave to the reader.","fun_headline_variants_meta":{"raw":{"variants":["Wall-crossing reveals number of Feynman master integrals","Exponential periods align Feynman integrals with wall-crossing","Thimble decomposition matches Feynman master integral expansion","Wall-crossing structure counts independent Feynman periods"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000628,"raw_usage":{"total_tokens":2950,"prompt_tokens":1039,"completion_tokens":1911,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":655,"completion_tokens_details":{"reasoning_tokens":1844}},"tokens_in":655,"tokens_out":1911,"duration_ms":17398,"temperature":1.0,"reasoning_tokens":1844,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T11:07:35.718386+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the kernel of the differential $d^{1}$ in the group cohomology of the nilmanifold boundary for the elliptic example: if the twisted cohomology of that boundary is not 0 ⊕ $C^{2}$ ⊕ $C^{2}$ ⊕ 0, then the middle cohomology in equation (4.74) would not be $C^{2}$ and the match with two master integrals would fail. A more global test would be to apply the cohomological count to a known one-loop Baikov family, such as the two-point integral, and check that the number of independent periods equals the number of master integrals obtained by integration by parts.","supporting_citations":[{"cited_title":"Feynman Integrals and Intersection Theory.JHEP, 02:139, 2019","cited_arxiv_id":null,"evidence_quote":"Establishes the identification of Feynman integrals with periods and intersection theory, the background against which master-integral decomposition is recast cohomologically."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the Baikov representation of Feynman integrals, the starting point for writing the integrand as e^{-γ log B}."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the review of the Baikov representation and integration by parts that motivates counting master integrals."},{"cited_title":"Cacciatori, Fabrizio Canfora, and Federica Muscolino","cited_arxiv_id":null,"evidence_quote":"Supplies the physical quartic-exponential integral from a nuclear-matter partition function used as the first worked example."},{"cited_title":"On classification of N=2 supersymmetric theories.Commun","cited_arxiv_id":null,"evidence_quote":"Introduces the wall-crossing structure in supersymmetric theories that the exponential-integral Stokes automorphisms generalize."}],"review_version":1}