Pith. sign in

REVIEW 3 major objections 3 minor 94 references

Differential Space of Feynman Integrals: Annihilators and $\mathcal{D}$-module

T0 review · 3 major / 3 minor · reviewed 2026-08-07 · deepseek-v4-flash

Pith's one-line read The paper builds annihilating differential operators for Feynman-like integrals directly from their parametric form, and conjectures that the resulting D-module's rank always equals the number of master integrals.

desk verdict A genuinely new two-term annihilator construction with honest conjecture, but a sign error in condition (C) and an unanalyzed boundary term leave the central proof incomplete as written. read the letter →

arxiv 2506.10456 v1 pith:6HHMZAHN submitted 2025-06-12 hep-th hep-phmath-phmath.MP

classification hep-thhep-phmath-phmath.MP
keywords FeynmanintegralstwistedperiodGriffiths-DworkreductionD-modulesholonomicrankPicard-FuchsoperatorsMacaulaymatrixmethodmaster
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

Feynman integrals in their parametric representation are twisted period integrals: the integral of a differential form times a multi-valued twist $u=U^\kappa/F^\eta$. The paper presents an algorithm that builds, directly from that representation, differential operators in the external variables that annihilate the integral, without first deriving integration-by-parts identities. Each annihilator has a simple form, the sum of two derivative operators of adjacent orders, and is constructed by a twisted Griffiths-Dwork reduction with boundary conditions that eliminate surface terms. Applied to hypergeometric functions, one-loop four-point integrals, the three-loop banana graph, and an AdS$_4$ Witten diagram, the method yields a $\mathcal{D}$-module whose holonomic rank coincides in every case with $|\chi(V)|$, the dimension of the twisted de Rham cohomology group, hence with the number of master integrals. The paper proposes this equality as a conjecture for generic Feynman-like integrals.

What carries the argument

The load-bearing object is the two-term annihilator $A_p=D_p+D_{p-1}$, a sum of two partial differential operators of orders $p$ and $p-1$ with respect to the external variables, designed so that $A_p\pi(s)=\int_\Gamma d\gamma=0$. The machinery is the twisted Griffiths-Dwork reduction formula, $d\gamma=(\eta+k-1)(\sum_i\lambda_i\partial_iF)/F^k\,\Omega-\sum_i\nabla_i\lambda_i/F^{k-1}\,\Omega$, together with the three conditions (A)-(C) that ensure $\gamma$ is a total derivative and vanishes on the boundary of $\Gamma$. The polynomials $\lambda_i=(q_i/q_0)\alpha_i$ are fixed by solving the syzygy equation $\sum_i q_i\theta_iF=q_0N_p$, where $\theta_i=\alpha_i\partial_i$; this simultaneously enforces the boundary and Jacobian-ideal conditions. The Macaulay matrix method then turns the annihilating ideal into a linear system whose row reduction yields the standard monomials, i.e. the independent derivative monomials that survive modulo the ideal; their number is the holonomic rank of the $\mathcal{D}$-module.

What would settle it

Find a Feynman-like twisted period integral — for instance with $\nu_i>1$ or with $F$ vanishing on a boundary component of the simplex — where the algorithm's operators give $\int_\Gamma d\gamma\neq 0$ because a boundary term from the twist survives, or find a generic example where the holonomic rank of the ideal generated by the $A_p$'s differs from $|\chi(V)|$ (equivalently from $\dim H^{n-1}_{dR}$). A direct check would be to compute both sides independently for a two-loop diagram at non-integer $d$ and compare.

Watch

Extended reading notes

Core claim

The paper's central assertion is that for a dimensionally regulated Feynman integral written as $\pi(s)=\int_\Gamma u\,\varphi$ with unit propagator powers, one can construct annihilators $A_p=D_p+D_{p-1}$ such that $A_p\pi(s)=\int_\Gamma d\gamma=\int_{\partial\Gamma}\gamma=0$, where $\gamma$ is a rational differential form built from polynomials $\lambda_i=(q_i/q_0)\alpha_i$ that satisfy three simultaneous conditions: (A) $\lambda_i=0$ when $\alpha_i=0$, so $\gamma$ vanishes on the coordinate faces of the simplex; (B) the numerator $N_p=\sum_i\lambda_i\partial_iF$ lies in the Jacobian ideal of the second Symanzik polynomial; and (C) $N_{p-1}=\sum_i\nabla_i\lambda_i$ with the twisted covariant derivative $\nabla_i=\partial_i+\partial_i\log U^\kappa$. These annihilators generate a left ideal in the rational Weyl algebra; the Macaulay matrix method on this ideal produces a set of standard monomials whose cardinality is the holonomic rank $r$. The paper's conjecture is that, under genericity and for sufficiently large $p$, $r=|\chi(V)|$, where $V$ is the zero locus of $U(\alpha)F(\alpha,s)$ in the torus, and by inclusion-exclusion this equals $\dim H^{n-1}_{dR}$, the number of master integrals. The equality is verified in all applications, and the resulting Picard-Fuchs operators, including a new one for the one-loop AdS$_4$ Witten diagram, match known integration-by-parts results where those are available.

Load-bearing premise

The proof that $A_p\pi(s)=0$ rests on the identity $\int_\Gamma d\gamma=\int_{\partial\Gamma}\gamma=0$; condition (A) makes $\lambda_i$ vanish on the coordinate faces $\alpha_i=0$, and the paper assumes no other boundary contributions arise from the multi-valued twist $u=U^\kappa/F^\eta$ — singular where $U=0$ or $F=0$ — or from the non-integer exponents of dimensional regularization, so if such contributions exist the constructed operators do not actually annihilate the integral.

Editorial extensions

If this is right

  • If Conjecture 1 holds, the number of master integrals of a Feynman-like integral is read off directly from the Euler characteristic $|\chi(V)|$, without an integration-by-parts reduction.
  • The annihilators produce Pfaffian systems and Picard-Fuchs operators directly from the parametric integrand, with the singular locus of the $\mathcal{D}$-module reproducing the Landau singularities of first and second type in the tested Feynman examples.
  • The construction works for generic (non-integer) dimensional-regularization parameters, and by design avoids inhomogeneous surface terms; operators built this way give the full differential system, not just the maximal-cut homogeneous part.
  • In the AdS$_4$ example the method yields the first exact-in-$\epsilon$ Picard-Fuchs operator for the one-loop Witten diagram, and in the $\epsilon\to0$ limit its factorization into linear factors matches the polylogarithmic form of the direct integration.

Reading between the lines

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

  • One natural test beyond the paper is to apply the construction to integrals with propagator powers $\nu_i>1$ or to graphs whose Symanzik polynomial $F$ vanishes on a face of the integration simplex; the paper states that non-unit powers can be treated along the same lines but gives no example, and the boundary condition (A) only controls the coordinate faces.
  • The rank equality suggests that the $\mathcal{D}$-module generated by external-variable annihilators may be an algebraic avatar of the twisted de Rham cohomology itself, not merely a source of differential equations; if that is so, the standard monomials should carry enough information to reconstruct the intersection-number decomposition of Feynman integrals.
  • Because $|\chi(V)|$ is in known cases the maximal likelihood degree of the associated toric variety, the conjecture hints at a toric-geometric proof strategy, analogous to the GKZ mixed-volume formula, for restricted period integrals.
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.

Referee Report

3 major / 3 minor

Summary. The paper proposes an algorithm for constructing differential operators in the external variables that annihilate Feynman-like twisted period integrals in parametric representation. The construction is based on a twisted Griffiths–Dwork reduction: for a candidate operator A_p = D_p + D_{p-1}, the integrand of A_p π(s) is written as an exact form dγ, and conditions (A), (B), and (C) are imposed so that γ vanishes on the boundary of the integration domain. The authors then use the Macaulay matrix method to obtain standard monomials, Pfaffian systems, and Picard–Fuchs operators, and they formulate Conjecture 1, which identifies the holonomic rank of the generated D-module with |χ(V)| and with the dimension of the twisted de Rham cohomology group. Applications include one-loop massless and one-mass four-point integrals, the three-loop equal-mass banana integral, a one-loop Witten diagram in AdS4, and hypergeometric 2F1 and 3F2 integrals. In each example, the computed rank agrees with known master-integral counts and the singular loci agree with Landau singularities.

Significance. If the construction is valid at the stated level of generality, the paper would provide a systematic, parametric-representation route from twisted period integrals to D-modules, Pfaffian systems, and Picard–Fuchs operators, complementing IBP-based methods and extending GKZ-type results to restricted integrals. The manuscript contains a substantial amount of explicit algebraic output: the generators of annihilating ideals and Picard–Fuchs operators are displayed for each example, and the agreement with independently known IBP/PF results and with Landau singularities is a genuine strength. The proposed rank–cohomology conjecture is interesting and testable. However, the central proof that A_p annihilates π(s) is not closed as written, and the completeness of the generated ideal rests on a heuristic step. These issues affect the degree of confidence one can place in the main claims.

major comments (3)
  1. [Griffiths–Dwork reduction and annihilators, Eqs. (8), (11), (12)] Equation (8), also derived in the Supplementary Material as Eqs. (39)–(41), gives dγ = (η+k−1)(Σ_i λ_i ∂_i F)/F^k Ω − (Σ_i ∇_i λ_i)/F^{k−1} Ω. Setting k=p and comparing with Eq. (5), the required conditions are N_p = (η+p−1)Σ_i λ_i ∂_i F and N_{p−1} = −Σ_i ∇_i λ_i. The paper instead states N_p = Σ_i λ_i ∂_i F and N_{p−1} = Σ_i ∇_i λ_i. The prefactor (η+p−1) can be absorbed into a redefinition of the λ_i, but the sign in Eq. (12) cannot be absorbed. As written, Eq. (6) does not follow from the stated conditions. This is a load-bearing point for the construction of annihilators, and it needs to be corrected or explicitly resolved by a different convention for ∇_i or N_{p−1}; the examples should then be rechecked under the corrected convention.
  2. [Setup and Griffiths–Dwork reduction, Eqs. (6)–(10)] The proof that A_p π(s)=0 rests on the identity ∫_Γ dγ = ∫_{∂Γ} γ and on Condition (A). The form γ in Eq. (7) contains the twist u = U^κ/F^η, which is multivalued and, in the dimensionally regulated cases considered, has non-integral exponents κ and η that can be negative, e.g. the three-loop banana has κ = 4−2d, which is −2+4ε at d = 4−2ε. The form also has poles on F=0. No regularization or branch-cut analysis is supplied to justify that the boundary integral vanishes in the dimensionally regulated sense. Condition (A) controls the behaviour of λ_i at the coordinate faces, but it does not by itself control γ on F=0, on U=0, or the behaviour of products λ_i u at faces when the exponents are non-integral. Since the equality ∫_Γ dγ = ∫_{∂Γ} γ is the mechanism by which the constructed operators are annihilators, this gap is central. The example checks are reassuring, but they do not replace an argument at the level of generality claimed in the paper.
  3. [Algorithm, 'Annihilators' paragraph, and Conjecture 1] The completeness of the annihilating ideal used for the rank computation is asserted through a heuristic coefficient choice. The text states that after Condition (C) leaves independent coefficients c_{I,ind}, 'operationally, we found that setting all but one of the c_{I,ind} to a non-zero value, at once, and going through the entire set of the independent coefficients produces the sufficient set of representatives of I.' No proof or algorithmic criterion is given that this produces a complete generating set for the annihilating ideal, and Conjecture 1 similarly refers to 'sufficiently large p' without a bound or stopping rule. Since the holonomic rank r=|Std| is then compared with dim H^{n-1}_{dR} in all examples, the empirical rank equality is only as strong as this completeness assumption. This should be either proved in the stated class of integrals, restricted by an explicit genericity condition, or clearly separated from the theorem-like statements in the applications.
minor comments (3)
  1. [Conjecture 1 and Eq. (16)] Equation (16) identifies r=|χ(V)| with |n−χ(V_E(f))| = dim H^{n-1}_{dR}. Since χ(V) need not be positive, the use of the absolute value deserves a brief explanation; otherwise the sign conventions in the Euler characteristic computation are ambiguous.
  2. [Setup, sentence after Eq. (2)] The paper restricts to ν_i=1, with footnote [68] promising that ν_i>1 can be treated 'along the same lines.' This restriction is significant for the claimed generality and should be stated in the abstract or introduction rather than only in the body.
  3. [Applications, Witten integral paragraph] In the Witten example, the statement that the PF operator is 'derived here for the first time' and that annihilation was 'verified up to this order' mixes an exact derivation with a truncated/order-by-order check. The main text should make clear which statements are exact in ε and which are verified only in a finite order expansion.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: annihilators are built from an algebraic total-derivative matching, and the rank–cohomology equality is an independently tested conjecture, not a fitted input.

full rationale

The derivation of the annihilators is a direct construction: Ap pi is matched, via the twisted Griffiths-Dwork identity (eqs. (6)-(8)) and conditions (A)-(C), to the integral of an explicit total derivative dgamma that vanishes at the boundary under condition (A). The differential operators are solved from syzygy and covariant-derivative constraints rather than assumed to be the known IBP/PF equations, and the resulting Ap pi = 0 is a sufficiency statement of the construction, not an input. The holonomic rank is then computed from the Macaulay matrix built out of these operators and compared, in each application, with independently evaluated Euler characteristics |chi(V)| and, where available, with IBP/PF operators and direct integration results; the general equality is explicitly labelled Conjecture 1, with the paper stating that sufficient conditions are not yet known. Citations to the authors' prior work supply algorithmic tools (Macaulay matrix, restricted Pfaffian systems, Euler characteristic relations) that are reproducible or externally checkable mathematical facts, so they are not load-bearing circular support. No fitted parameter is relabelled as a prediction. The internal sign and regularization issues that a reader might locate in the Stokes argument concern correctness of the proof as written, not circularity: they do not make the argument equivalent to its inputs by definition.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central derivation relies on standard D-module and Euler-characteristic machinery plus three paper-specific choices: the unit-exponent restriction, the boundary-condition ansatz, and the completeness heuristic for selecting independent coefficients. No genuinely new physical entities are introduced; the proposed rank-cohomology equivalence is a conjecture, not an input.

free parameters (3)
  • order p of annihilators = chosen per example (3, 2, 4, 2)
    The construction requires 'p sufficiently large' but no general bound is stated; in each example the order is chosen by hand and the rank depends on it.
  • Macaulay degree l = not specified
    The rank and standard monomial basis in eq. (20) depend on a chosen maximal derivative degree l; no automatic stopping rule is given.
  • choice of independent coefficients c_{I,ind} = all but one set to nonzero values
    The set of annihilator representatives that generate the ideal is selected by this heuristic from the Algorithm section; the resulting ideal depends on the choice.
assumptions (4)
  • domain assumption Twisted Griffiths-Dwork formula (eq. 8) and its use in Stokes theorem hold for the multi-valued twist u = U^kappa / F^eta on the simplex Gamma without anomalous boundary contributions.
    Eq. (6) equates Ap pi(s) to integral over the boundary of Gamma of gamma; this requires the boundary integral to vanish exactly. The paper ensures lambda_i = 0 on coordinate faces but does not analyze singularities of u on U = 0 or F = 0.
  • ad hoc to paper Unit denominator exponents nu_i = 1 are sufficient to derive D-modules for the considered integrals.
    Stated in Setup and footnote [68] as an empirical finding; no proof is given, and all applications use nu_i = 1.
  • standard math Euler characteristic and inclusion-exclusion identity (16) correctly compute dim H^{n-1}_{dR} for the twisted and relative twisted cohomologies.
    Used in Conjecture 1 and eq. (16), following [14, 62, 64, 75].
  • ad hoc to paper Conjecture 1: under genericity and for large p, the holonomic rank of the generated ideal equals |chi(V)|.
    The paper states this is not proven and has only been verified for toy models and the cases shown.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Differential Space of Feynman Integrals: Annihilators and $\mathcal{D}$-module." pith.science (2026). https://pith.science/paper/6HHMZAHN

@misc{pith2026250610456,
  author       = {Pith},
  title        = {Pith review of: Differential Space of Feynman Integrals: Annihilators and $\mathcalD$-module},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6HHMZAHN}},
  note         = {Machine review of arXiv:2506.10456}
}
abstract

We present a novel algorithm for constructing differential operators with respect to external variables that annihilate Feynman-like integrals and give rise to the associated $\mathcal{D}$-modules, based on Griffiths-Dwork reduction. By leveraging the Macaulay matrix method, we derive corresponding relations among partial differential operators, including systems of Pfaffian equations and Picard-Fuchs operators. Our computational approach is applicable to twisted period integrals in projective coordinates, and we showcase its application to Feynman graphs and Witten diagrams. The method yields annihilators and their algebraic relations for generic regulator values, explicitly avoiding contributions from surface terms. In the cases examined, we observe that the holonomic rank of the $\mathcal{D}$-modules coincides with the dimension of the corresponding de Rham co-homology groups, indicating an equivalence relation between them, which we propose as a conjecture.

Figures

Figures reproduced from arXiv: 2506.10456 by the authors.

Figure 2
Figure 2. One-loop 1-mass 4-point graph. Kinematic vari [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 1
Figure 1. One-loop massless 4-point graph. Kinematic vari [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 3
Figure 3. Three-loop banana graph. Kinematic variables: [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: One-loop s-channel Witten diagram in AdS4. Finally, we consider the integral corresponding to the s-channel one-loop graph in fig. 4 studied in [86], hereby homogenized w.r.t. external variables. • Symanzik polynomials: U1 =(α2 + α3 + α4 + α5)α1 + (α2 + α3 + α4)α5 , (3…

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

94 extracted references · 36 canonical work pages

  1. [1]

    Differential equations method: New technique for massive Feynman diagrams calculation,

    A. V. Kotikov, “Differential equations method: New technique for massive Feynman diagrams calculation,” Phys. Lett. B254, 158–164 (1991)

  2. [2]

    Differential equation method: The Cal- culation of N point Feynman diagrams,

    A. V. Kotikov, “Differential equation method: The Cal- culation of N point Feynman diagrams,” Phys. Lett. B 267, 123–127 (1991), [Erratum: Phys.Lett.B 295, 409– 409 (1992)]

  3. [3]

    Differential equations for Feynman graph amplitudes,

    Ettore Remiddi, “Differential equations for Feynman graph amplitudes,” Nuovo Cim. A110, 1435–1452 (1997), arXiv:hep-th/9711188

  4. [4]

    Differential equations for two loop four point functions,

    T. Gehrmann and E. Remiddi, “Differential equations for two loop four point functions,” Nucl. Phys. B580, 485–518 (2000), arXiv:hep-ph/9912329

  5. [5]

    Feynman Dia- grams and Differential Equations,

    Mario Argeri and Pierpaolo Mastrolia, “Feynman Dia- grams and Differential Equations,” Int. J. Mod. Phys. A 22, 4375–4436 (2007), arXiv:0707.4037 [hep-ph]

  6. [6]

    Multiloop integrals in dimensional regularization made simple,

    Johannes M. Henn, “Multiloop integrals in dimensional regularization made simple,” Phys. Rev. Lett.110, 251601 (2013), arXiv:1304.1806 [hep-th]

  7. [7]

    Magnus and Dyson Series for Master Integrals,

    Mario Argeri, Stefano Di Vita, Pierpaolo Mastrolia, Edoardo Mirabella, Johannes Schlenk, Ulrich Schubert, and Lorenzo Tancredi, “Magnus and Dyson Series for Master Integrals,” JHEP03, 082 (2014), arXiv:1401.2979 [hep-ph]

  8. [8]

    Simplified differential equa- tions approach for Master Integrals,

    Costas G. Papadopoulos, “Simplified differential equa- tions approach for Master Integrals,” JHEP07, 088 (2014), arXiv:1401.6057 [hep-ph]

Show all 94 references
  1. [9]

    A Theorem on Analytical Calculability of Four Loop Renormalization Group Functions,

    F. V. Tkachov, “A Theorem on Analytical Calculability of Four Loop Renormalization Group Functions,” Phys. Lett.B100, 65–68 (1981)

  2. [10]

    Integration by Parts: The Algorithm to Calculate beta Functions in 4 Loops,

    K. G. Chetyrkin and F. V. Tkachov, “Integration by Parts: The Algorithm to Calculate beta Functions in 4 Loops,” Nucl. Phys.B192, 159–204 (1981)

  3. [11]

    High precision calculation of multiloop Feynman integrals by difference equations,

    S. Laporta, “High precision calculation of multiloop Feynman integrals by difference equations,” Int. J. Mod. Phys. A15, 5087–5159 (2000), arXiv:hep-ph/0102033

  4. [12]

    Feynman In- tegrals and Intersection Theory,

    Pierpaolo Mastrolia and Sebastian Mizera, “Feynman In- tegrals and Intersection Theory,” JHEP02, 139 (2019), arXiv:1810.03818 [hep-th]

  5. [13]

    Decomposition of Feynman In- tegrals on the Maximal Cut by Intersection Numbers,

    Hjalte Frellesvig, Federico Gasparotto, Stefano Laporta, Manoj K. Mandal, Pierpaolo Mastrolia, Luca Mattiazzi, and Sebastian Mizera, “Decomposition of Feynman In- tegrals on the Maximal Cut by Intersection Numbers,” JHEP05, 153 (2019), arXiv:1901.11510 [hep-ph]

  6. [14]

    Vector Space of Feynman Integrals and Mul- tivariate Intersection Numbers,

    Hjalte Frellesvig, Federico Gasparotto, Manoj K. Man- dal, Pierpaolo Mastrolia, Luca Mattiazzi, and Sebastian Mizera, “Vector Space of Feynman Integrals and Mul- tivariate Intersection Numbers,” Phys. Rev. Lett.123, 201602 (2019), arXiv:1907.02000 [hep-th]

  7. [15]

    First studies of differential equations for Feynman inte- grals, earlier than the introduction of the dimensional regularization, also made use of the parametric represen- tation [16–18], and more recently in [19]

  8. [16]

    Differential prop- erties of Feynman amplitudes,

    B. Jacksic V. de Alfaro and T. Regge, “Differential prop- erties of Feynman amplitudes,” inHigh Energy Physics and Elementary Particles, Proceedings of Lectures at the International Centre for Theoretical Physics, Trieste (1965) p. 263

  9. [17]

    Differential properties of Feynman ampli- tudes. 1

    M. Giffon, “Differential properties of Feynman ampli- tudes. 1.” Nuovo Cim. A61, 663–684 (1969)

  10. [18]

    Differential equations for one-loop generalized feynman integrals,

    G. Barucchi and G. Ponzano, “Differential equations for one-loop generalized feynman integrals,” Journal of Mathematical Physics14, 396–401 (1973)

  11. [19]

    Di- mensionally regulated pentagon integrals,

    Zvi Bern, Lance J. Dixon, and David A. Kosower, “Di- mensionally regulated pentagon integrals,” Nucl. Phys. B412, 751–816 (1994), arXiv:hep-ph/9306240

  12. [20]

    On the periods of certain rational integrals: I,

    Philip A. Griffiths, “On the periods of certain rational integrals: I,” Annals of Mathematics90, 460–495 (1969)

  13. [21]

    On the zeta function of a hypersur- face,

    Bernard Dwork, “On the zeta function of a hypersur- face,” Publications Math´ ematiques de l’IH´ES12, 5–68 (1962)

  14. [22]

    On the zeta function of a hypersurface: Ii,

    Bernard Dwork, “On the zeta function of a hypersurface: Ii,” Annals of Mathematics80, 227–299 (1964)

  15. [23]

    Picard-Fuchs equations for Feynman inte- grals,

    Stefan M¨ uller-Stach, Stefan Weinzierl, and Raphael Zayadeh, “Picard-Fuchs equations for Feynman inte- grals,” Commun. Math. Phys.326, 237–249 (2014), arXiv:1212.4389 [hep-ph]

  16. [24]

    Differential equations on unitarity cut sur- faces,

    Mao Zeng, “Differential equations on unitarity cut sur- faces,” JHEP06, 121 (2017), arXiv:1702.02355 [hep-th]

  17. [25]

    Feynman integrals in dimensional regularization and extensions of Calabi- Yau motives,

    Kilian B¨ onisch, Claude Duhr, Fabian Fischbach, Al- brecht Klemm, and Christoph Nega, “Feynman integrals in dimensional regularization and extensions of Calabi- Yau motives,” JHEP09, 156 (2022), arXiv:2108.05310 [hep-th]

  18. [26]

    Algorithms for mini- mal Picard–Fuchs operators of Feynman integrals,

    Pierre Lairez and Pierre Vanhove, “Algorithms for mini- mal Picard–Fuchs operators of Feynman integrals,” Lett. Math. Phys.113, 37 (2023), arXiv:2209.10962 [hep-th]

  19. [27]

    Integration- by-parts identities and differential equations for parametrised Feynman integrals,

    Daniele Artico and Lorenzo Magnea, “Integration- by-parts identities and differential equations for parametrised Feynman integrals,” JHEP03, 096 (2024), arXiv:2310.03939 [hep-ph]

  20. [28]

    Algorithm for differential equations for Feynman integrals in gen- 10 eral dimensions,

    Leonardo de la Cruz and Pierre Vanhove, “Algorithm for differential equations for Feynman integrals in gen- 10 eral dimensions,” Lett. Math. Phys.114, 89 (2024), arXiv:2401.09908 [hep-th]

  21. [29]

    Macaulay matrix for Feynman integrals: linear rela- tions and intersection numbers,

    Vsevolod Chestnov, Federico Gasparotto, Manoj K. Mandal, Pierpaolo Mastrolia, Saiei J. Matsubara- Heo, Henrik J. Munch, and Nobuki Takayama, “Macaulay matrix for Feynman integrals: linear rela- tions and intersection numbers,” JHEP09, 187 (2022), arXiv:2204.12983 [hep-th]

  22. [30]

    Restrictions of Pfaffian systems for Feynman integrals,

    Vsevolod Chestnov, Saiei J. Matsubara-Heo, Henrik J. Munch, and Nobuki Takayama, “Restrictions of Pfaffian systems for Feynman integrals,” JHEP11, 202 (2023), arXiv:2305.01585 [hep-th]

  23. [31]

    D-module techniques for solving differential equations in the context of Feyn- man integrals,

    Johannes Henn, Elizabeth Pratt, Anna-Laura Sattel- berger, and Simone Zoia, “D-module techniques for solving differential equations in the context of Feyn- man integrals,” Lett. Math. Phys.114, 87 (2024), arXiv:2303.11105 [hep-th]

  24. [32]

    Differen- tial equations for moving hyperplane arrangements,

    Ana¨ elle Pfister and Anna-Laura Sattelberger, “Differen- tial equations for moving hyperplane arrangements,” (2025), arXiv:2412.09479 [math.CO]

  25. [33]

    The associatedD- module is the space of all differential operatorsDmodulo those that annihilateπ(s)

    Annihilators of a given functionπ(s) form an ideal: ifA annihilatesπ(s), that isAπ(s) = 0, then left multipli- cation with any other differential operator will do so too and thus belong to the ideal as well. The associatedD- module is the space of all differential operatorsDmo...

  26. [34]

    Hypergeometric functions and toral mani- folds,

    Izrail Moiseevich Gelfand, A. V. Zelevinskii, and Mikhail Kapranov, “Hypergeometric functions and toral mani- folds,” Functional Analysis and Its Applications23, 94– 106 (1989)

  27. [35]

    Gen- eralized euler integrals and a-hypergeometric functions,

    I.M Gelfand, M.M Kapranov, and A.V Zelevinsky, “Gen- eralized euler integrals and a-hypergeometric functions,” Advances in Mathematics84, 255–271 (1990)

  28. [36]

    For a list of applications of the GKZ approach to Feyn- man integrals, see [37–58]

  29. [37]

    Periods of Feynman Dia- grams and GKZ D-Modules,

    Emad Nasrollahpoursamami, “Periods of Feynman Dia- grams and GKZ D-Modules,” , arXiv:1605.04970 (2016), arXiv:1605.04970 [math-ph]

  30. [38]

    Feynman integrals as A- hypergeometric functions,

    Leonardo de la Cruz, “Feynman integrals as A- hypergeometric functions,” (2019), arXiv:1907.00507 [math-ph]

  31. [39]

    Kinematic singularities of Feyn- man integrals and principal A-determinants,

    Ren´ e Pascal Klausen, “Kinematic singularities of Feyn- man integrals and principal A-determinants,” JHEP02, 004 (2022), arXiv:2109.07584 [hep-th]

  32. [40]

    Hypergeometric Series Represen- tations of Feynman Integrals by GKZ Hypergeometric Systems,

    Ren´ e Pascal Klausen, “Hypergeometric Series Represen- tations of Feynman Integrals by GKZ Hypergeometric Systems,” JHEP04, 121 (2020), arXiv:1910.08651 [hep- th]

  33. [41]

    thesis, Mainz U

    Ren´ e Pascal Klausen,Hypergeometric Feynman integrals, Ph.D. thesis, Mainz U. (2023), arXiv:2302.13184 [hep-th]

  34. [42]

    Hypergeometric Functions and Feynman Diagrams,

    Mikhail Kalmykov, Vladimir Bytev, Bernd A. Kniehl, Sven-Olaf Moch, Bennie F. L. Ward, and Scott A. Yost, “Hypergeometric Functions and Feynman Diagrams,” in Antidifferentiation and the Calculation of Feynman Am- plitudes(2020) arXiv:2012.14492 [hep-th]

  35. [43]

    Cohen-Macaulay Property of Feynman Integrals,

    Felix Tellander and Martin Helmer, “Cohen-Macaulay Property of Feynman Integrals,” Commun. Math. Phys. 399, 1021–1037 (2023), arXiv:2108.01410 [hep-th]

  36. [44]

    Conformal integrals in four dimensions,

    Aritra Pal and Koushik Ray, “Conformal integrals in four dimensions,” JHEP10, 087 (2022), arXiv:2109.09379 [hep-th]

  37. [45]

    Conformal integrals in various dimensions and Clifford groups,

    Aritra Pal and Koushik Ray, “Conformal integrals in various dimensions and Clifford groups,” (2023), arXiv:2303.17326 [hep-th]

  38. [46]

    FeynGKZ: A Mathematica package for solving Feynman integrals using GKZ hypergeometric systems,

    B. Ananthanarayan, Sumit Banik, Souvik Bera, and Sudeepan Datta, “FeynGKZ: A Mathematica package for solving Feynman integrals using GKZ hypergeometric systems,” Comput. Phys. Commun.287, 108699 (2023), arXiv:2211.01285 [hep-th]

  39. [47]

    Feynman integrals of Grassmannians,

    Tai-Fu Feng, Hai-Bin Zhang, and Chao-Hsi Chang, “Feynman integrals of Grassmannians,” Phys. Rev. D 106, 116025 (2022), arXiv:2206.04224 [hep-th]

  40. [48]

    GKZ-system of the 2-loop self energy with 4 prop- agators,

    Tai-Fu Feng, Hai-Bin Zhang, Yan-Qing Dong, and Yang Zhou, “GKZ-system of the 2-loop self energy with 4 prop- agators,” (2022), arXiv:2209.15194 [hep-th]

  41. [49]

    GKZ-hypergeometric systems for Feyn- man integrals,

    Tai-Fu Feng, Chao-Hsi Chang, Jian-Bin Chen, and Hai-Bin Zhang, “GKZ-hypergeometric systems for Feyn- man integrals,” Nucl. Phys. B953, 114952 (2020), arXiv:1912.01726 [hep-th]

  42. [50]

    GKZ hypergeometric systems of the three-loop vacuum Feynman integrals,

    Hai-Bin Zhang and Tai-Fu Feng, “GKZ hypergeometric systems of the three-loop vacuum Feynman integrals,” (2023), arXiv:2303.02795 [hep-th]

  43. [51]

    On feynman graphs, matroids, and gkz- systems,

    Uli Walther, “On feynman graphs, matroids, and gkz- systems,” Letters in Mathematical Physics112, 120 (2022)

  44. [52]

    Symbol Alphabets from the Landau Singular Locus,

    Christoph Dlapa, Martin Helmer, Georgios Papathana- siou, and Felix Tellander, “Symbol Alphabets from the Landau Singular Locus,” (2023), arXiv:2304.02629 [hep- th]

  45. [53]

    Vector spaces of generalized Euler integrals,

    Daniele Agostini, Claudia Fevola, Anna-Laura Sattel- berger, and Simon Telen, “Vector spaces of generalized Euler integrals,” Commun. Num. Theor. Phys.18, 327– 370 (2024), arXiv:2208.08967 [math.AG]

  46. [54]

    Feynman Integral Relations from GKZ Hypergeometric Systems,

    Henrik J. Munch, “Feynman Integral Relations from GKZ Hypergeometric Systems,” PoSLL2022, 042 (2022), arXiv:2207.09780 [hep-th]

  47. [55]

    Thel-loop Banana Amplitude from GKZ Systems and relative Calabi-Yau Periods,

    Albrecht Klemm, Christoph Nega, and Reza Safari, “Thel-loop Banana Amplitude from GKZ Systems and relative Calabi-Yau Periods,” JHEP04, 088 (2020), arXiv:1912.06201 [hep-th]

  48. [56]

    Analytic structure of all loop banana integrals,

    Kilian B¨ onisch, Fabian Fischbach, Albrecht Klemm, Christoph Nega, and Reza Safari, “Analytic structure of all loop banana integrals,” JHEP05, 066 (2021), arXiv:2008.10574 [hep-th]

  49. [57]

    Polytope symmetries of Feyn- man integrals,

    Leonardo de la Cruz, “Polytope symmetries of Feyn- man integrals,” Phys. Lett. B854, 138744 (2024), arXiv:2404.03564 [hep-th]

  50. [58]

    Reductions of GKZ systems and applications to cosmological corre- lators,

    Thomas W. Grimm and Arno Hoefnagels, “Reductions of GKZ systems and applications to cosmological corre- lators,” JHEP04, 196 (2025), arXiv:2409.13815 [hep-th]

  51. [59]

    Differential equations for the Feynman integral of the self-energy diagram,

    Valentina A. Golubeva, “Differential equations for the Feynman integral of the self-energy diagram,” Differ. Uravn.9, 1298–1309 (1973)

  52. [60]

    The differential equations for the feynman amplitude of a single-loop graph with four vertices,

    V. A. Golubeva and V. Z. ´Enol’skii, “The differential equations for the feynman amplitude of a single-loop graph with four vertices,” Mathematical notes of the Academy of Sciences of the USSR23, 63–66 (1978)

  53. [61]

    integration vari- ables were earlier considered in [62]

    Annihilators of Feynman integrals w.r.t. integration vari- ables were earlier considered in [62]

  54. [62]

    Feynman integral relations from para- metric annihilators,

    Thomas Bitoun, Christian Bogner, Rene Pascal Klausen, and Erik Panzer, “Feynman integral relations from para- metric annihilators,” Lett. Math. Phys.109, 497–564 (2019), arXiv:1712.09215 [hep-th]

  55. [63]

    Critical points and number of master integrals,

    Roman N. Lee and Andrei A. Pomeransky, “Critical points and number of master integrals,” JHEP11, 165 (2013), arXiv:1308.6676 [hep-ph]

  56. [64]

    Twisted cohomology and likelihood ideals,

    Saiei-Jaeyeong Matsubara-Heo and Simon Telen, 11 “Twisted cohomology and likelihood ideals,” Advances in Applied Mathematics165, 102832 (2025)

  57. [65]

    Hypergeometric functions and rings generated by monomials,

    Alan Adolphson, “Hypergeometric functions and rings generated by monomials,” Duke Math. J.73, 269–290 (1994)

  58. [66]

    6 (Springer, Berlin, 2000) pp

    Mutsumi Saito, Bernd Sturmfels, and Nobuki Takayama, Gr¨ obner deformations of hypergeometric differential equations, Algorithms and computation in mathematics, Vol. 6 (Springer, Berlin, 2000) pp. viiI, 254

  59. [67]

    The maximum likelihood degree,

    Fabrizio Catanese, Serkan Hosten, Amit Khetan, and Bernd Sturmfels, “The maximum likelihood degree,” American Journal of Mathematics128, 671–697 (2006)

  60. [68]

    Applications to integrals withN∋ν i >1 can be derived along the same lines

  61. [69]

    Relations among the rational Weyl algebraR m, the polynomial Weyl algebra D, and their sheaf-theoretic counterparts are also dis- cussed in these references

    See, e.g., [70–72] for a review ofD-modules theory and related computational techniques. Relations among the rational Weyl algebraR m, the polynomial Weyl algebra D, and their sheaf-theoretic counterparts are also dis- cussed in these references

  62. [70]

    (Springer, Tokyo, Japan, 2014)

    Takayuki Hibiet al.,Gr¨ obner bases, 2013th ed. (Springer, Tokyo, Japan, 2014)

  63. [71]

    Algorithms forD-modules—restriction, tensor product, localization, and local cohomology groups,

    Toshinori Oaku and Nobuki Takayama, “Algorithms forD-modules—restriction, tensor product, localization, and local cohomology groups,” J. Pure Appl. Algebra 156, 267–308 (2001)

  64. [72]

    Mutsumi Saito, Bernd Sturmfels, and Nobuki Takayama, Gr¨ obner deformations of hypergeometric differential equations, Algorithms and computation in mathematics (Springer, Berlin, Germany, 2011)

  65. [73]

    , sm]⟨∂1,

    The theory ofD-modules is developed for the polynomial Weyl algebraD=C[s, . . . , sm]⟨∂1, . . . , ∂m⟩⊊R m with the same commutation relation (14) as in the rational Weyl algebraR m. The correspondence between the two setups comes from the observation, that for anyIinR m there ...

  66. [74]

    The Chern-Schwartz-Macpherson class of an embeddable scheme,

    Paolo Aluffi, “The Chern-Schwartz-Macpherson class of an embeddable scheme,” Forum of Mathematics, Sigma 7, e30 (2019)

  67. [75]

    Status of Intersection Theory and Feynman Integrals,

    Sebastian Mizera, “Status of Intersection Theory and Feynman Integrals,” PoSMA2019, 016 (2019), arXiv:2002.10476 [hep-th]

  68. [76]

    Intersection theory for twisted cohomologies and twisted Riemann’s period re- lations I,

    Koji Cho and Keiji Matsumoto, “Intersection theory for twisted cohomologies and twisted Riemann’s period re- lations I,” Nagoya Math. J.139, 67–86 (1995)

  69. [77]

    Relative Twisted Homology and Co- homology Groups Associated with Lauricella’s FD,

    Keiji Matsumoto, “Relative Twisted Homology and Co- homology Groups Associated with Lauricella’s FD,” Funkcialaj Ekvacioj67, 105–147 (2024)

  70. [78]

    (16) the following relation is employed: χ({α1 · · ·αn ̸= 0} \V) =χ({α 1 · · ·αn ̸= 0})−χ(V) =−χ(V) =χ(P n−1)−χ(V E(f)) =n−χ(V E(f)),with χ({α1 · · ·αn ̸= 0}) = 0 andχ(P n−1) =n

    In eq. (16) the following relation is employed: χ({α1 · · ·αn ̸= 0} \V) =χ({α 1 · · ·αn ̸= 0})−χ(V) =−χ(V) =χ(P n−1)−χ(V E(f)) =n−χ(V E(f)),with χ({α1 · · ·αn ̸= 0}) = 0 andχ(P n−1) =n. For further details on the correspondence between the Euler char- acteristic and the dimens...

  71. [79]

    From the momentum space representation, assuming standard quadratic denominators, the degree of homo- geneity isℓd/2− P i νi, beingν i the denominator powers

  72. [80]

    Fr´ ed´ eric Chyzak,The ABC of Creative Telescoping — Algorithms, Bounds, Complexity, Accreditation to super- vise research, Ecole Polytechnique X (2014)

  73. [81]

    Computing periods of rational integrals,

    Pierre Lairez, “Computing periods of rational integrals,” Mathematics of Computation85, 1719–1752 (2015)

  74. [82]

    Faster multivariate integration in d-modules,

    Hadrien Brochet, Fr´ ed´ eric Chyzak, and Pierre Lairez, “Faster multivariate integration in d-modules,” (2025), arXiv:2504.12724 [cs.SC]

  75. [83]

    The solutionsq i in (18) of the inhomogeneous syzygy equation (19) are defined up to solutionsq i,H of the ho- mogeneous syzygy equation P i qi,H θiF = 0. In our al- gorithm, this freedom is exploited for the determination ofλ i, satisfying the A, B, C conditions, to build ann...

  76. [84]

    The solution of syzygy equations, the holonomic rank and the singular locus of the generatedD-modules are computed within the computer algebra systemSingular [90]

    Our algorithm is implemented inMathematica[89]. The solution of syzygy equations, the holonomic rank and the singular locus of the generatedD-modules are computed within the computer algebra systemSingular [90]. Euler characteristic of the zero locus of affine vari- eties are ...

  77. [85]

    The three-loop equal-mass banana integral inε-factorised form with meromorphic modular forms,

    Sebastian P¨ ogel, Xing Wang, and Stefan Weinzierl, “The three-loop equal-mass banana integral inε-factorised form with meromorphic modular forms,” JHEP09, 062 (2022), arXiv:2207.12893 [hep-th]

  78. [86]

    Analytical evaluation of AdS 4 Witten diagrams as flat space multi-loop Feynman integrals,

    Till Heckelbacher, Ivo Sachs, Evgeny Skvortsov, and Pierre Vanhove, “Analytical evaluation of AdS 4 Witten diagrams as flat space multi-loop Feynman integrals,” JHEP08, 052 (2022), arXiv:2201.09626 [hep-th]

  79. [87]

    Hypergeometric Dis- criminants,

    Saiei-Jaeyeong Matsubara-Heo, “Hypergeometric Dis- criminants,” (2025), arXiv:2505.13163 [math.AG]

  80. [88]

    Algebraic study of systems of partial differential equations,

    Masaki Kashiwara, “Algebraic study of systems of partial differential equations,” M´ em. Soc. Math. France (N.S.) , xiv+72 (1995), translated by Jean-Pierre Shneiders and Andrea D’Agnolo

  81. [89]

    Mathematica, Version 14.2,

    Wolfram Research, Inc., “Mathematica, Version 14.2,” Champaign, IL, 2024

  82. [90]

    Singular4-4-0 — A computer algebra system for polynomial computations,

    Wolfram Decker, Gert-Martin Greuel, Gerhard Pfister, and Hans Sch¨ onemann, “Singular4-4-0 — A computer algebra system for polynomial computations,”http:// www.singular.uni-kl.de(2024)

  83. [91]

    Macaulay2, a software system for research in algebraic geometry,

    Daniel R. Grayson and Michael E. Stillman, “Macaulay2, a software system for research in algebraic geometry,” Available athttp://www.math.uiuc.edu/Macaulay2/

  84. [92]

    Openxm project (including risa/asir distribution),

    “Openxm project (including risa/asir distribution),” http://www.openxm.org

  85. [93]

    FiniteFlow: multivariate functional re- construction using finite fields and dataflow graphs,

    Tiziano Peraro, “FiniteFlow: multivariate functional re- construction using finite fields and dataflow graphs,” JHEP07, 031 (2019), arXiv:1905.08019 [hep-ph]

  86. [94]

    Maple 2024.1. Maplesoft, a division of Waterloo Maple Inc., Waterloo, Ontario,

    “Maple 2024.1. Maplesoft, a division of Waterloo Maple Inc., Waterloo, Ontario,”

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.