Pith. sign in

REVIEW 1 major objections 9 minor 1 cited by

This paper proves that the position-space and parameter-space integrals associated to the same graph are equal, and that they evaluate to single-valued multiple zeta values.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · deepseek-v4-flash

2026-08-01 01:57 UTC pith:FMCPKCPD

load-bearing objection The equality theorem is the real result and looks essentially sound; the single-valued MZV evaluation is a sketch leaning on external technology, and that part should be read as conditional until the hypotheses are checked. the 1 major comments →

arxiv 2607.25595 v1 pith:FMCPKCPD submitted 2026-07-28 math.QA math-phmath.MPmath.NT

Graph integrals, Feynman periods, and single-valued multiple zeta values

Jean-Luc Portner This is my paper
classification math.QA math-phmath.MPmath.NT MSC 11M3281Q3014H10
keywords canonical integralsRW-integralssingle-valued multiple zeta valuesFeynman periodsgraph complexesdeformation quantisationDodgson polynomialstropical Torelli map
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

For graphs with n+1 vertices and 2n edges, two entirely different-looking integrals are claimed to be the same number: the RW-integral, taken over vertex positions in the complex plane, and the canonical integral, taken over positive edge parameters. The paper proves this equality by constructing an auxiliary integral over both position and parameter space and integrating it in the two possible orders. From this, it derives that every such graph integral is a single-valued multiple zeta value of weight n, and that every single-valued multiple zeta value appears as a rational linear combination of Feynman periods. This matters because it ties together the stable cohomology of general linear groups, deformation quantisation, and the arithmetic of Feynman integrals.

Core claim

For any real 2n×n matrix A of rank n with no zero row, the RW-integral (a configuration-space integral of logarithms and arguments of the linear forms a_i(z)) equals the canonical integral (the simplex integral of the invariant form tr((L(x)^{-1}dL(x))^{2n-1}) with L(x)=A^T diag(x)A). For graphs this says that the position-space and parameter-space pictures of the same Feynman-type object coincide. The equality implies that canonical integrals of graphs with E=2V−2 are single-valued multiple zeta values, that every single-valued multiple zeta value is a rational linear combination of Feynman periods, and that the associated RW and canonical cocycles in the commutative graph complex agree.

What carries the argument

The auxiliary integral I_aux over C^n × σ_{2n}, whose two orders of integration reproduce the two sides of the equality. One direction uses a regularised Feynman parametrisation and a partial-fraction identity to pass to the RW-integral; the other uses complex Gaussian integration, a differential-operator expansion into permanents, and a determinant-permanent matrix identity to recover the canonical form. The new explicit formula for the canonical form, with denominator Ψ^n instead of the naively expected Ψ^{2n−1}, is a by-product of this machinery.

Load-bearing premise

The claim that RW-integrals evaluate to single-valued multiple zeta values rests on an external integration theorem, sketched but not proved in this paper; if that theorem fails or requires additional normalisation, the number-theoretic conclusions weaken even though the equality of the integrals might still hold.

What would settle it

Compute I_can(G) and I_RW(G) for any graph with E=2V−2 and check equality to high precision; a single mismatch would falsify the main theorem. More specifically, evaluate the five-wheel graph: the paper predicts both equal 1260 ζ(5), so any deviation from that number, or any non-single-valued component at weight 5, would falsify Proposition 1.3.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • Every canonical graph integral for E=2V−2 graphs is a single-valued multiple zeta value of weight n, so the coincidences observed in examples are systematic.
  • The space of single-valued multiple zeta values is contained in the space of Feynman periods; each such number has a concrete graph-theoretic realisation as a rational combination of convergent Feynman integrals.
  • The RW and canonical constructions define the same degree-zero cocycle on the commutative graph complex, so cohomology classes computed by either method are identical.
  • The proof yields a new explicit formula for canonical forms that is more efficient for computation and explains the cancellations that had made direct calculation difficult.
  • The analytic mechanism explains the observed weight drop in canonical integrals: their Hodge weight is 2n, rather than the generic 4n−6 bound for Feynman integrals.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Because RW-integrals have explicit integration algorithms, the equality turns canonical integrals into effectively computable numbers; one could generate tables of Feynman periods for small graphs and probe new relations among single-valued multiple zeta values.
  • The paper notes that products of canonical forms can involve non-single-valued values (such as ζ(3,5) for K_6); a natural test is whether any generalised RW-integral could reproduce such products, or whether the equality is special to the primitive class.
  • The inclusion of single-valued multiple zeta values in Feynman periods suggests a possible interaction with the coaction principle for Feynman periods; one could test whether the coaction of a canonical integral stays within the subalgebra generated by canonical integrals.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

1 major / 9 minor

Summary. The paper proves an equality between two families of graph integrals for graphs with n+1 vertices and 2n edges (equivalently 2n×n matrices of rank n): the canonical integral I_can(A), defined by integrating the trace form β^{2n-1}_L over the positive simplex, and the RW integral I_RW(A), defined as a configuration-space integral over C^n modulo scaling. Theorem 2.2 (and Theorem 1.1 in the graph case) states I_RW(A)=I_can(A). The proof introduces an auxiliary integral over C^n×σ_{2n}, regularized by truncation and a counterterm, and evaluates it in two orders: Feynman/Schwinger parametrization plus scale integration gives the RW integral (Section 5), and Gaussian integration plus a determinant–permanent combinatorial identity gives the canonical integral (Sections 6–7). The paper also claims that for graphs these integrals evaluate to single-valued multiple zeta values of weight n (Prop. 1.3/3.6, by external single-valued integration technology), that RW-integrals generate Z^sv as a Q-algebra (Prop. 1.5, citing Rossi–Willwacher and Brown), and hence Z^sv⊆P_Feyn (Theorem 1.8). In addition, the paper gives an explicit formula for canonical forms (Theorem 7.1), a reduction of the RW integral to a single logarithm term (Prop. 1.13), and corollaries about cocycles in the commutative graph complex and values of the Voronoi-cochain I_can.

Significance. If the proof is correct, the main equality theorem is a substantial bridge between two a priori independent integral representations: the parameter-space canonical integrals (which are connected to GL_n cohomology and tropical Torelli maps) and position-space RW integrals from deformation quantization. The proof is not circular and does not rely on fitted parameters; the core equality is supported by a detailed analytic argument with absolute-convergence statements (Thm 4.3), a regularized Feynman parametrization (Thm 5.2), dominated convergence bounds (Prop. 5.5, Lemma 6.5), and a purely combinatorial matrix identity (Thm 7.4). A further consequence, that all single-valued MZVs lie in the space of Feynman periods (Thm 1.8), follows from the equality plus established results provided the generation statement (Prop. 1.5) is accepted. The main weakness is that the number-theoretic evaluation of the integrals (Prop. 1.3 / Prop. 3.6) is only sketched, with the weight-raising primitive-and-residue step resting on external single-valued integration results [BPP20; VZ22b]; this is load-bearing for the headline 'evaluate to single-valued MZVs' claim. The equality theorem itself appears to

major comments (1)
  1. [Prop. 3.6 continuation] The paper's headline number-theoretic conclusion—that canonical/RW integrals of graphs are single-valued MZVs of weight n (Prop. 1.3, Cor. 2.4)—rests entirely on this sketch. The sketch asserts that the RW-integrand lies in U^{2n-2}_1(X_n) and that each successive integration raises the polylogarithmic weight by exactly 1, with no boundary terms and no normalization changes. This is a nontrivial analytic statement: the integrand mixes log(|z_i-z_j|^2) prefactors with dlog(|z_e|^2) forms, and the cited primitive formula must apply to this exact form after the reduction of Prop. 3.5, including the signs and the (2πi)^{-(n-1)} normalization. The manuscript does not verify the hypotheses of [BPP20; VZ22b] point by point. In particular, the claim that evaluations at 0 vanish is asserted without proof, and the weight-raising step is stated rather than demonstrated. This is load-bearing for the
minor comments (9)
  1. [§1, definition of I_RW(G)] The displayed formula for I_RW(G) in the introduction is typographically awkward: the double sum over l,d and the wedge over e≠l,d should be clearer; also the reality statement for odd/even n should be phrased more precisely.
  2. [Remark 1.2] The comparison factor with [RW14] is stated without derivation. Please provide a short justification or a precise reference.
  3. [§3, Prop. 3.5 proof] The replacement of darg-forms by dlog-forms uses a standard but nontrivial mixed-term argument; the sign calculation would benefit from one extra sentence explaining why only the mixed terms survive and why the sign is (-1)^{n-1}.
  4. [§4, proof of Thm 4.1] The auxiliary integrand is written with 'a_e da_e + a_e da_e' in eq. (12), while the RW integrand later uses 'a_e da_e - a_e da_e'. This sign convention is confusing; please state the equivalence explicitly.
  5. [Thm 5.2] In the regularized Feynman parametrization, the direction of the map φ is stated incorrectly: φ maps σ^ε_n to σ^0_n, not the reverse. Please fix.
  6. [Lemma 6.5] The constants in the bound are inconsistent: the hypothesis says |p_i|^2<ε^2/(8n) but the proof uses ε^2/(16n); also a parenthesis appears to be missing in the final displayed inequality.
  7. [§7.1, Prop. 7.8] The sign cancellation in the involution proof is described verbally; writing out the signs of the sorting permutations would improve rigor and readability.
  8. [Appendix A, Lemma A.10] The argument x^{-1} of L appears inconsistently: the statement uses L(x)^{-1} in one place and L(x^{-1}) in the proof. Please harmonize.
  9. [General] The manuscript contains many typos and formatting artifacts ('ALUED', 'Is', missing spaces, inconsistent notation for dlog(|a_e|^2)). These should be cleaned up before publication.

Circularity Check

2 steps flagged

No significant circularity: the equality theorem is proven through an independently constructed auxiliary integral, the single-valuedness step is a sketched use of external technology (a rigor gap), and the sole self-citation [Por23] is non-load-bearing.

specific steps
  1. other [Section 1.1, wheel-graph example, eq. (3)]
    "More generally, the RW-integrals and the canonical integrals have been computed for all wheel graphs in [BS25; Por23]."

    This is the paper's only self-citation. It is not load-bearing: it only corroborates the illustrative wheel-graph values, jointly with the independent [BS25] (Brown–Schnetz), and no theorem proof cites [Por23]. Flagged to document the self-citation; no circular reduction occurs.

  2. other [Proposition 3.6, Sketch of proof]
    "The proof sketch is based on the theory of [BPP20; VZ22b]."

    This step imports the single-valued primitive-and-residue integration calculus on which Proposition 1.3's single-valued-MZV conclusion rests, but only as a sketch. The concern is completeness, not circularity: the cited theory is external, developed for deformation quantization and closed-string amplitudes, and does not assume the present equality or values. Whether it applies verbatim to the RW-integrand with Proposition 3.5's prefactors is asserted rather than proven; the main equality theorem (Theorem 2.2) does not depend on this step.

full rationale

The central derivation I_RW(A) = I_can(A) (Theorems 1.1/2.2) is self-contained and non-circular. The proof constructs an auxiliary integral I_aux on C^n × σ_{2n}; Theorem 5.1 evaluates it by Feynman parametrization and partial fractions to the RW-integral, Theorem 6.1 evaluates it by complex Gaussian integration to a determinant-permanent expression, and Theorem 7.1 (with the combinatorial identity of Theorem 7.4 and Proposition 3.11 citing [Bro21]) identifies that expression with the canonical form β^{2n-1}_L. The RW and canonical integrals have independent definitions (Def. 3.1 and Def. 3.9), and the equality is derived by computing the common ancestor in two orders. No fitted parameters appear anywhere, and no step assumes the target result. The value statements form a separate layer: Prop. 1.3 = Theorem 1.1 + Prop. 3.6, where Prop. 3.6 is explicitly labeled 'Sketch of proof' and invokes the independently published single-valued integration technology of [BPP20; VZ22b; BD21; SS19]. Whether that external calculus applies verbatim to this integrand (weight/degree bookkeeping, absence of boundary terms) is asserted rather than demonstrated; this is a correctness/rigor gap, not circularity, because the cited theory was developed for other objects (deformation quantization, closed-string amplitudes) and does not presuppose the present results. Prop. 1.5 rests on [RW14] + [Bro14, Lemma 5.1] (the KZ/anti-KZ associator ratio generating Z^sv), and Theorem 1.8 additionally uses closure of P_Feyn ([Bro09, Prop. 40]; [Pan20, Thm. 2]) — all independent external publications. The only self-citation is [Por23], used together with independent [BS25] purely for illustrative wheel-graph values; it is not load-bearing. External benchmarks (W_3 = 60ζ(3), eq. (2)–(3)) corroborate the statements. Verdict: no circular step; the score of 2 reflects only the one benign, non-load-bearing self-citation, while the Prop. 3.6 sketch is flagged as a completeness risk, not as circularity.

Axiom & Free-Parameter Ledger

0 free parameters · 8 axioms · 0 invented entities

Central claims rely on previously established theorems rather than free numerical parameters. The proof imports Brown's canonical-form structure, Rossi–Willwacher/Brown generation of svMZVs, single-valued integration technology, and closure of Feynman periods under multiplication. No invented entities or fitted constants are introduced.

axioms (8)
  • domain assumption Brown's theorem: canonical form β^{2n-1}_L has the form P_G(x)/Ψ^{(n+1)/2} Ω(x), with the Dodgson-polynomial trace formula [Bro21, Thm 2.1, Cor 6.24].
    Invoked in Proposition 3.11 and Theorem 7.1; not re-proved in this paper, and used to identify the auxiliary-integral output with the canonical form.
  • domain assumption RW integrals generate the algebra of single-valued multiple zeta values via the KZ/anti-KZ associator [RW14; Bro14, Lemma 5.1].
    Used in Proposition 1.5 and Proposition 3.6; this is the bridge from RW-integral values to Zsv and is externally referenced rather than derived here.
  • domain assumption Single-valued integration technology: iterated integration of single-valued hyperlogarithms with residues yields single-valued MZVs [BPP20; VZ22b; Sch15].
    Used in the proof sketch of Proposition 3.6 to conclude IRW(G)∈Zsv_n; this is the least developed part of the present text.
  • domain assumption Canonical integrals are Feynman periods [Bro21].
    Needed for Corollary 1.7 and Theorem 1.8 to place RW integrals and then Zsv inside PFeyn.
  • domain assumption The space PFeyn of Feynman periods is closed under multiplication via the two-vertex join [Bro09, Prop 40] or face maps [Pan20, Thm 2].
    Used in Theorem 1.8 to pass from algebra generators to Q-linear combinations of Feynman periods.
  • domain assumption Kontsevich vanishing lemma for angular forms [Kon03, Section 6.6.1].
    Used in Remark 3.4 and Proposition 3.7 to prove invariance and vanishing of RW integrals.
  • standard math Standard measure-theoretic exchange theorems (Fubini, dominated convergence, Leibniz integral rule).
    Used throughout Sections 4–6 to justify interchanging limits, derivatives, and integrals in the proof of IRW=Ican.
  • domain assumption Voronoi complex computes locally finite homology of P_n/GL_n(Z) [EGS10; Bro25].
    Used in Corollary 2.4 to describe values of the canonical cocycle on the Voronoi complex.

pith-pipeline@v1.3.0-alltime-deepseek · 44648 in / 12305 out tokens · 112203 ms · 2026-08-01T01:57:54.465050+00:00 · methodology

0 comments
read the original abstract

The Borel classes generating the stable cohomology of the general linear group can be represented by invariant differential forms. It is known that pulling these forms back along a tropical Torelli map yields canonical convergent integrals associated to graphs, which are closely connected to the cohomology of $\mathrm{GL}_n$ and of graph complexes. A natural question is what numbers these graph integrals are. We answer this for primitive canonical integrals by showing that they coincide with a family of complex position-space integrals arising in deformation quantisation. As a consequence, canonical integrals of graphs evaluate to single-valued multiple zeta values. We further deduce that every single-valued multiple zeta value occurs as a rational linear combination of Feynman periods of graphs with massless propagators. Finally, in the commutative graph complex, our result implies that the two associated cocycles agree.

Figures

Figures reproduced from arXiv: 2607.25595 by Jean-Luc Portner.

Figure 1
Figure 1. Figure 1: The three-wheel graph W3 with labeled vertices and labeled directed edges on the left and the general wheel graph Wn on the right. the left. For instance, the first row encodes that edge 1 is oriented from vertex 1 to vertex 2, as depicted in figure 1. Removing any column from ε yields a reduced incidence matrix, denoted εe, and the corresponding reduced Laplacian matrix is given by LG = εe T diag(x1, . . … view at source ↗
Figure 2
Figure 2. Figure 2: The three-wheel graph G with external legs on the left, and the two-loop master graph Ge on the right [BW03; Gro12]. Example 1.6. Consider the three-wheel graph with four external legs G, shown on the left of figure 2. In its scalar Feynman integral in dimensional regularisation, D = 4 − 2ε, the Feynman period appears as the pole in ε: IG({m2 i }, {p 2 i }, D) = 1 3ε Z σG Ω(x) Ψ2 + O(ε 0 ). On the other ha… view at source ↗
Figure 3
Figure 3. Figure 3: The auxiliary integral Iaux relates the RW and the canonical integral. Let n ≥ 3, and let d ∈ [2n]. We write z = (z1, . . . , zn) for the coordinates on C n . The simplex σ2n is initially parametrised by coordinates α = (α1, . . . , α2n). Later, when passing from the auxiliary integral to the canonical integral, we change variables by setting αi = x −1 i . Thus the auxiliary integral is written in the coor… view at source ↗
Figure 4
Figure 4. Figure 4: Graphs with distinct cycle structure arising from different σ and ϱ. In the first case the graph decomposes into two cycles: an odd cycle containing l, and an even cycle entirely contained in S ⊔ T. In contrast, the graph on the right consists of a single cycle of odd length. The cancellations in eq. (27) are characterised as follows: Proposition 7.8. The sum of all terms Wd(S, T, l, σ, ϱ) in eq. (27) for … view at source ↗
Figure 5
Figure 5. Figure 5: Passing from (S, T) to (S ′ , T′ ) by advancing the cycle (158732) by one. Let α, β ∈ Sn−1 be the permutations that sort S ′ and T ′ increasingly and extend them to permutations on [n] by fixing n. Then define new sets and permutations Sb = α(S ′ ), Tb = β(T ′ ) as well as ϱb := β −1 ◦ ϱ ◦ α −1 , σb = α −1 ◦ σ ◦ β −1 . This defines a map f(S, T, l, σ, ϱ) = (S, b T , l, b σ, b ϱb). To see that f is an invol… view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. The motivic Lie algebra embeds into the cohomology of the general linear group

    math.AG 2026-07 conditional novelty 7.0

    The motivic Lie algebra of mixed Tate motives over Z is shown to embed canonically into unstable compactly-supported cohomology of GL_g(Z)-locally symmetric spaces and of A_g, via canonical graph cocycles that are gen...

Reference graph

Works this paper leans on

52 extracted references · 25 canonical work pages · cited by 1 Pith paper · 3 internal anchors

  1. [1]

    Axelrod, Scott and Singer, I. M. , date =. Chern-. doi:10.4310/jdg/1214454681 , url =

    arXiv

  2. [2]

    Baikov, P. A. and Chetyrkin, K. G. , date =. Four Loop Massless Propagators:. doi:10.1016/j.nuclphysb.2010.05.004 , url =

    work page doi:10.1016/j.nuclphysb.2010.05.004 2010

  3. [3]

    doi:10.1007/JHEP05(2025)120 , url =

    Combinatorial Proof of a Non-Renormalization Theorem , author =. doi:10.1007/JHEP05(2025)120 , url =

    doi:10.1007/jhep05(2025)120 2025

  4. [4]

    doi:10.1007/s00222-020-00970-x , url =

    Multiple Zeta Values in Deformation Quantization , author =. doi:10.1007/s00222-020-00970-x , url =

    work page doi:10.1007/s00222-020-00970-x

  5. [5]

    Wonderful

    Berghoff, Marko , date =. Wonderful. doi:10.4310/CNTP.2015.v9.n3.a2 , url =. 1411.5583 , eprinttype =

    Pith/arXiv arXiv 2015

  6. [6]

    doi:10.1140/epjc/s2003-01389-7 , url =

    The Massless Two-Loop Two-Point Function , author =. doi:10.1140/epjc/s2003-01389-7 , url =

    work page doi:10.1140/epjc/s2003-01389-7

  7. [7]

    Bloch, Spencer and Esnault, Hélène and Kreimer, Dirk , date =. On. doi:10.1007/s00220-006-0040-2 , url =

    doi:10.1007/s00220-006-0040-2

  8. [8]

    doi:10.1016/j.crma.2004.02.001 , url =

    Polylogarithmes Multiples Uniformes En Une Variable , author =. doi:10.1016/j.crma.2004.02.001 , url =

    work page doi:10.1016/j.crma.2004.02.001 2004

  9. [9]

    , author =

    Single-Valued Hyperlogarithms and Unipotent Differential Equations. , author =

  10. [10]

    Brown, Francis C. S. , date =. On the Periods of Some. doi:10.48550/arXiv.0910.0114 , url =. 0910.0114 , eprinttype =

    Pith/arXiv arXiv doi:10.48550/arxiv.0910.0114

  11. [11]

    Spanning Forest Polynomials and the Transcendental Weight of

    Brown, Francis and Yeats, Karen , date =. Spanning Forest Polynomials and the Transcendental Weight of. doi:10.1007/s00220-010-1145-1 , url =. 0910.5429 , eprinttype =

    Pith/arXiv arXiv

  12. [12]

    Brown, Francis and Schnetz, Oliver , date =. A. doi:10.1215/00127094-1644201 , url =. 1006.4064 , eprinttype =

    Pith/arXiv arXiv

  13. [13]

    doi:10.1017/fms.2014.18 , url =

    Single-Valued Motivic Periods and Multiple Zeta Values , author =. doi:10.1017/fms.2014.18 , url =

    work page doi:10.1017/fms.2014.18 2014

  14. [14]

    Periods and

    Brown, Francis , date =. Periods and. doi:10.48550/arXiv.1512.09265 , url =. 1512.09265 , eprinttype =

    Pith/arXiv arXiv doi:10.48550/arxiv.1512.09265

  15. [15]

    Feynman Amplitudes, Coaction Principle, and Cosmic

    Brown, Francis , date =. Feynman Amplitudes, Coaction Principle, and Cosmic. doi:10.4310/CNTP.2017.v11.n3.a1 , url =

    doi:10.4310/cntp.2017.v11.n3.a1 2017

  16. [16]

    Invariant

    Brown, Francis , date =. Invariant. doi:10.3842/SIGMA.2021.103 , url =

    work page doi:10.3842/sigma.2021.103 2021

  17. [17]

    Brown, Francis and Dupont, Clément , date =. Single-. doi:10.1007/s00220-021-03969-4 , url =

    work page doi:10.1007/s00220-021-03969-4

  18. [18]

    Bordifications of the Moduli Spaces of Tropical Curves and Abelian Varieties, and Unstable Cohomology of

    Brown, Francis , date =. Bordifications of the Moduli Spaces of Tropical Curves and Abelian Varieties, and Unstable Cohomology of. doi:10.1007/s00222-025-01335-y , url =

    work page doi:10.1007/s00222-025-01335-y

  19. [19]

    The Wheel Classes in the Locally Finite Homology of

    Brown, Francis and Schnetz, Oliver , date =. The Wheel Classes in the Locally Finite Homology of. doi:10.2140/gt.2025.29.4389 , url =

    work page doi:10.2140/gt.2025.29.4389 2025

  20. [20]

    doi:10.1134/S1063779618050118 , url =

    Poisson Brackets Symmetry from the Pentagon-Wheel Cocycle in the Graph Complex , author =. doi:10.1134/S1063779618050118 , url =. 1712.05259 , eprinttype =

    Pith/arXiv arXiv

  21. [21]

    , date =

    Buring, Ricardo and Kiselev, Arthemy V. , date =. Universal Cocycles and the Graph Complex Action on Homogeneous. doi:10.1134/S1547477120050088 , url =. 1912.12664 , eprinttype =

    Pith/arXiv arXiv 1912

  22. [22]

    doi:10.1090/jams/965 , url =

    Tropical Curves, Graph Complexes, and Top Weight Cohomology of M_g , author =. doi:10.1090/jams/965 , url =. 1805.10186 , eprinttype =

    Pith/arXiv arXiv

  23. [23]

    Combinatorics of the Tropical

    Chan, Melody , date =. Combinatorics of the Tropical. doi:10.2140/ant.2012.6.1133 , url =

    work page doi:10.2140/ant.2012.6.1133 2012

  24. [24]

    Chetyrkin, K. G. and Kataev, A. L. and Tkachov, F. V. , date =. New Approach to Evaluation of Multiloop. doi:10.1016/0550-3213(80)90289-8 , url =

    work page doi:10.1016/0550-3213(80)90289-8

  25. [25]

    On a Theorem of

    Conant, James and Vogtmann, Karen , date =. On a Theorem of. doi:10.2140/agt.2003.3.1167 , url =. math/0208169 , eprinttype =

    Pith/arXiv arXiv 2003

  26. [26]

    doi:10.1007/BF01589496 , url =

    Wonderful Models of Subspace Arrangements , author =. doi:10.1007/BF01589496 , url =

    doi:10.1007/bf01589496

  27. [27]

    Gillam, W D , langid =

  28. [28]

    doi:10.4310/CNTP.2019.v13.n4.a1 , url =

    Dodgson Polynomial Identities , author =. doi:10.4310/CNTP.2019.v13.n4.a1 , url =

    work page doi:10.4310/cntp.2019.v13.n4.a1 2019

  29. [29]

    Grozin, A. G. , date =. Massless Two-Loop Self-Energy Diagram:. doi:10.1142/S0217751X12300189 , url =. 1206.2572 , eprinttype =

    Pith/arXiv arXiv

  30. [30]

    Interpolating Polynomials and Divided Differences , author =

  31. [31]

    Relating depth graded and block graded motivic Lie algebras

    Keilthy, Adam , date =. Relating Depth Graded and Block Graded Motivic. doi:10.48550/arXiv.2307.08089 , url =. 2307.08089 , eprinttype =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2307.08089

  32. [32]

    Deformation

    Kontsevich, Maxim , date =. Deformation. doi:10.1023/B:MATH.0000027508.00421.bf , url =

    arXiv

  33. [33]

    doi:10.1017/CBO9781107340688 , url =

    Permanents , author =. doi:10.1017/CBO9781107340688 , url =

    work page doi:10.1017/cbo9781107340688

  34. [34]

    doi:10.1016/j.nuclphysb.2013.05.025 , url =

    On the Analytic Computation of Massless Propagators in Dimensional Regularization , author =. doi:10.1016/j.nuclphysb.2013.05.025 , url =

    work page doi:10.1016/j.nuclphysb.2013.05.025 2013

  35. [35]

    Panzer, Erik and Schnetz, Oliver , date =. The. doi:10.4310/CNTP.2017.v11.n3.a3 , url =

    doi:10.4310/cntp.2017.v11.n3.a3 2017

  36. [36]

    Panzer, Erik , editor =. Some. Periods in. doi:10.1007/978-3-030-37031-2_2 , abstract =

    work page doi:10.1007/978-3-030-37031-2_2

  37. [37]

    Computing the Single-Valued

    Portner, Jean-Luc , date =. Computing the Single-Valued

  38. [38]

    and Willwacher, Thomas , date =

    Rossi, Carlo A. and Willwacher, Thomas , date =. P. doi:10.48550/arXiv.1404.2047 , url =. 1404.2047 , eprinttype =

    Pith/arXiv arXiv doi:10.48550/arxiv.1404.2047 2047

  39. [39]

    Closed Strings as Single-Valued Open Strings: A Genus-Zero Derivation , shorttitle =

    Schlotterer, Oliver and Schnetz, Oliver , date =. Closed Strings as Single-Valued Open Strings: A Genus-Zero Derivation , shorttitle =. doi:10.1088/1751-8121/aaea14 , url =

    work page doi:10.1088/1751-8121/aaea14

  40. [40]

    Quantum Periods:

    Schnetz, Oliver , date =. Quantum Periods:. doi:10.4310/CNTP.2010.v4.n1.a1 , url =

    work page doi:10.4310/cntp.2010.v4.n1.a1 2010

  41. [41]

    doi:10.4310/CNTP.2014.v8.n4.a1 , url =

    Graphical Functions and Single-Valued Multiple Polylogarithms , author =. doi:10.4310/CNTP.2014.v8.n4.a1 , url =

    work page doi:10.4310/cntp.2014.v8.n4.a1 2014

  42. [42]

    doi:10.1103/PhysRevD.97.085018 , url =

    Numbers and Functions in Quantum Field Theory , author =. doi:10.1103/PhysRevD.97.085018 , url =

    doi:10.1103/physrevd.97.085018

  43. [43]

    , date =

    Stieberger, S. , date =. Closed. doi:10.1088/1751-8113/47/15/155401 , url =. 1310.3259 , eprinttype =

    Pith/arXiv arXiv

  44. [44]

    doi:10.4310/ATMP.2022.v26.n2.a5 , url =

    Closed String Amplitudes from Single-Valued Correlation Functions , author =. doi:10.4310/ATMP.2022.v26.n2.a5 , url =

    work page doi:10.4310/atmp.2022.v26.n2.a5 2022

  45. [45]

    doi:10.4310/ATMP.2022.v26.n2.a5 , url =

    Single-Valued Hyperlogarithms, Correlation Functions and Closed String Amplitudes , author =. doi:10.4310/ATMP.2022.v26.n2.a5 , url =. 1812.03018 , eprinttype =

    Pith/arXiv arXiv 2022

  46. [46]

    Willwacher, Thomas , date =. M. doi:10.1007/s00222-014-0528-x , url =. 1009.1654 , eprinttype =

    Pith/arXiv arXiv

  47. [47]

    doi:10.48550/arXiv.2508.13724 , url =

    The 11-Loop Graph Cohomology , author =. doi:10.48550/arXiv.2508.13724 , url =. 2508.13724 , eprinttype =

    Pith/arXiv arXiv doi:10.48550/arxiv.2508.13724

  48. [48]

    The triconnected Kontsevich graph complex

    Willwacher, Thomas , date =. The Triconnected. doi:10.48550/arXiv.2503.17131 , url =. 2503.17131 , eprinttype =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2503.17131

  49. [49]

    Perfect forms and the cohomology of modular groups

    Perfect Forms and the Cohomology of Modular Groups , author =. doi:10.48550/arXiv.1001.0789 , url =. 1001.0789 , eprinttype =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1001.0789

  50. [50]

    doi:10.1007/s00222-016-0647-7 , url =

    Logarithms and Deformation Quantization , author =. doi:10.1007/s00222-016-0647-7 , url =

    work page doi:10.1007/s00222-016-0647-7

  51. [51]

    doi:10.48550/arXiv.2312.17720 , url =

    Regularized Integrals and Manifolds with Log Corners , author =. doi:10.48550/arXiv.2312.17720 , url =. 2312.17720 , eprinttype =

    work page doi:10.48550/arxiv.2312.17720

  52. [52]

    doi:10.1215/00127094-2010-022 , url =

    Torelli Theorem for Graphs and Tropical Curves , author =. doi:10.1215/00127094-2010-022 , url =

    work page doi:10.1215/00127094-2010-022 2010