Pith. sign in

REVIEW 3 major objections 4 minor 28 references

DSSYK's low-temperature corrections are governed by modular Eisenstein series, not arbitrary asymptotics.

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

2026-08-02 06:46 UTC pith:CFE5ETDA

load-bearing objection A genuinely interesting DSSYK paper with solid core machinery and a real verification gap: the all-n closed forms (5.35) that carry the headline resummation are checked to n=20 and proven only for n=1. the 3 major comments →

arxiv 2607.11828 v2 pith:CFE5ETDA submitted 2026-07-13 hep-th

Modular structures in the DSSYK partition function

Matteo Beccaria , Eleonora Alfinito This is my paper
classification hep-th
keywords DSSYKpartition functionquasi-modular Eisenstein seriesS-dualitynon-perturbative correctionsdivisor sumstriangular numbersbilocal-Liouville saddles
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.

The paper studies the low-temperature (large-β) expansion of the exact disk partition function of double-scaled SYK (DSSYK) at fixed coupling λ. It claims that this expansion is not arbitrary: it is organized by the classical ring of quasi-modular Eisenstein series E₂, E₄, E₆ and the Ramanujan differential identities that close that ring. Using modular S-duality to pass to the semiclassical small-λ regime, the non-perturbative corrections controlled by exp(−4π²/λ) are determined in closed form up to second order in λ, and the whole series resums into a compact expression built from the same modular functions. The paper further shows that this entire structure follows from one exact differential equation coupling a modular derivative to temperature derivatives, and that the non-perturbative sector of Z itself is supported exactly on triangular exponents that match known bulk winding saddles. A sympathetic reader would care because the result replaces an asymptotic germ with number-theoretic scaffolding, making the low-temperature physics of DSSYK far more constrained and structurally transparent.

Core claim

Starting from the exact Bessel-function representation Z(β)=F(x) with x=2β/√(λ(1−q)), the paper shows that expanding at large x yields coefficients that are polynomials in E₂, E₄, E₆ times (q;q)∞³. After S-duality, log Z splits into a perturbative tower and a non-perturbative sector in q̃=e^{−4π²/λ}. At each order q̃ⁿ the correction is a genuine function of y=24π²/(λβ), with general closed forms involving the divisor sums σ₁(n) and σ₃(n). Summing over n resums into 3 log(Λ;Λ)∞ + (1−E₂(Λ)) y (1 + y/3π²) λ/96 + O(λ²), with Λ=e^{y/3} q̃. Finally, the paper proves that the non-perturbative sector of Z(β) itself is supported exactly, to all orders in λ, on triangular exponents n=j(j+1)/2, equal t

What carries the argument

The quasi-modular Eisenstein ring: E₂, E₄, E₆ with the Ramanujan identities D E₂=(E₂²−E₄)/12, D E₄=(E₂E₄−E₆)/3, D E₆=(E₂E₆−E₄²)/2, together with D(q;q)∞=(E₂−1)(q;q)∞/24. The Bessel coefficients are generated by applying the ladder operators 32D+3−4j(j−1) to S₀=(q;q)∞³; because D preserves this ring, every coefficient lies in it. The master mechanism is the exact differential equation 8DF=[x²∂²ₓ+3x∂ₓ−x²]F, first order in the modular derivative and second order in x, from which both the large-x quasi-modular expansion and the high-temperature analytic expansion follow.

Load-bearing premise

The closed formulas for f_{n,0}, f_{n,1}, and f_{n,2} for general n are load-bearing; they are stated after checking n up to 20, with only the n=1 case proven, so the whole divisor-sum resummation rests on that pattern holding for all n.

What would settle it

Compute the q̃ⁿ coefficient of the non-perturbative log Z in the small-λ expansion for some n>20 and compare with (5.35): if the predicted combination of σ₁(n) and σ₃(n) fails, the closed resummation is wrong at that order. Alternatively, derive the general-n coefficient directly from the differential equation (6.3) and check the e^{ny/3} dependence and divisor-function form.

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

If this is right

  • If correct, the low-temperature expansion of DSSYK is fully controlled by modular data and divisor sums, so all-order completions can be written in terms of finitely many modular functions.
  • The known leading non-perturbative result—β→∞, all orders in q̃: Z ∝ η(q̃)³—is a degenerate case of a genuine y-dependent structure, and the subleading β-dependence is universal rather than model-specific.
  • The exact differential equation gives a first-principles derivation of the persistent (1−q)^{n−1} factorization in the high-temperature expansion, generating any order without resumming Bessel series.
  • The support of the non-perturbative sector of Z on triangular numbers selects exactly the winding/conical-defect saddles, leaving no unexplained exponents.
  • Because the coefficients are divisor functions σ₁ and σ₃, the result ties DSSYK low-temperature physics to elementary number theory, so any bulk explanation must reproduce these arithmetic weights.

Where Pith is reading between the lines

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

  • The general-n closed forms (5.35) are inferred from n up to 20 and rigorously proven only for n=1; if the pattern breaks at some larger n, the resummation (5.38) fails at that q̃ⁿ order even though the leading β→∞ coefficients survive.
  • The exact differential equation may be a special case of a modular heat-kernel structure that extends to other observables, such as two-point functions and Krylov complexity; the paper itself points to this as a natural next step.
  • The S-duality technique used here, inverting the Eisenstein ring rather than manipulating Bessel functions, could organize the non-perturbative completion of other DSSYK quantities beyond the partition function.

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

3 major / 4 minor

Summary. The paper studies the low-temperature (large-β) expansion of the exact disk partition function Z(β) of double-scaled SYK at fixed λ, starting from the Bessel-function representation (2.7). It shows that the large-x expansion can be organized by the ring of quasi-modular Eisenstein series E₂, E₄, E₆ (eq. 4.22), and derives an exact differential equation (6.3) relating modular and temperature derivatives. Using modular S-duality, the paper splits log Z into a perturbative tower and a non-perturbative sector in q̃ = e^{-4π²/λ}, claiming closed-form corrections at each order in q̃ up to O(λ²) (eq. 5.35, 5.38), with coefficients built from divisor functions σ₁ and σ₃. It further claims that the non-perturbative sector of Z itself is exactly supported on triangular exponents n = j(j+1)/2, matching winding saddles of the bilocal-Liouville action from [25].

Significance. If the central claims hold, the paper reveals a remarkably clean modular/quasi-modular structure underlying DSSYK low-temperature physics, and provides the first explicit finite-y completion of the non-perturbative corrections beyond the β→∞ result of [23]. The paper has several genuine strengths: the exact differential equation (6.3) is a compact and elegant structural result; the derivation of the n=1 resummations in Appendix A is rigorous; the high-temperature recursion and divisibility proof in Section 6.3 is a nice first-principles derivation of known structure from [19]; and the benchmarks against [23], [28], [19] are explicitly verified. These give strong evidence that the formal modular manipulations are correct in spirit.

major comments (3)
  1. [§5.3, eq. (5.35)–(5.38)] The general-n closed forms for f_{n,0}, f_{n,1}, f_{n,2} are load-bearing for the headline resummation (5.38) and for the abstract's claim that non-perturbative corrections are 'determined in closed form' at each order in q̃. However, only the n=1 case is rigorously derived (Appendix A); for general n the paper reports a check up to n=20. The genuinely new y-dependent terms f_{n,1} and f_{n,2} are therefore not proven for n≥2. The paper itself acknowledges in Section 7 that the divisor-sum coefficients enter as boundary data fixed by matching, not derived from the differential equation. This is a verification gap rather than an observed contradiction, but it is load-bearing: if the pattern (5.35) breaks at any n>20, eq. (5.38) fails at the corresponding Λⁿ order. The authors should either prove (5.35) by induction (for example from (6.7) with boundary data from (4.22)) or explicitly labe
  2. [§5.3.1 (sparsity/triangular support)] The claim that the non-perturbative sector of Z(β) is exactly supported on triangular exponents for all orders in λ is not rigorously justified. The argument that D acts diagonally applies to the q-series before S-duality. After substituting the inversion rules (5.3)–(5.4), each φ_k(rq) becomes a polynomial in E₂(rq), E₄(rq), E₆(rq) times S₀(rq); such a product generically contains non-triangular powers of rq unless special identities hold. The two examples in (5.40) illustrate the phenomenon but do not constitute a proof for arbitrary k. Since this exact support is one of the paper's central claims (and is used in Section 7 for the saddle-spectrum match), a general proof or a more detailed inductive argument is needed.
  3. [§4.2, Remark after (4.24); §5] The derivation of the non-perturbative expansion (5.38) reorganizes the large-x asymptotic series (4.22), which has zero radius of convergence, via S-duality and then exponentiates log Z. The paper correctly notes in the Remark that lateral Borel corrections ~e^{-2x} are not studied. In the double-scaled regime y = 24π²/(λβ) fixed, one expects these corrections to be exponentially suppressed relative to q̃ = e^{-4π²/λ}, but this is not demonstrated. Because the central quantitative result (5.38) is obtained by formal manipulation of an asymptotic series, the paper should either provide the suppression argument or state explicitly that all non-perturbative claims are formal asymptotic statements (with the benchmarks as evidence).
minor comments (4)
  1. [Eq. (5.33)] In the display for f_{3,2}(y), 'eyy' should read 'e^y y' (the exponential is ambiguous).
  2. [Eq. (5.35)] The expression for f_{n,2}(y) has a parenthesis mismatch: the numerator of the 36π² term should be y(180 + 34 n y − 5 y σ₃(n)/σ₁(n)), not 'yp180...' as printed.
  3. [§5.1, eq. (5.16)] The relation between log β and b₂(β) is clear, but the notation b̄ᵢ vs bᵢ in (5.7)–(5.10) is easy to confuse; a short comment or consistent notation would help.
  4. [§4.2, end of Section 4] The remark about the asymptotic nature of (4.22) is valuable; it might be moved to the main text rather than a Remark to avoid readers missing the caveat.

Circularity Check

0 steps flagged

No circular derivation: the central NP resummation rests on an unproven all-n induction (n≤20), which is a verification gap, not a circular reduction; self-citations are non-load-bearing.

full rationale

The derivation chain is self-contained and externally benchmarked rather than circular. It starts from the exact Bessel/chord representations (2.7)/(2.5) taken from independent prior work [5,24]; the large-x quasi-modular expansion (4.22) follows from the Jacobi triple product and the Ramanujan identities, with the monomial-count check in §4.2 being a combinatorial consistency test, not an input. The small-λ S-duality reexpansion (5.3)–(5.4) is standard modular inversion. The non-perturbative functions f_{n,p}(y) for n=1 are rigorously derived in Appendix A; the general-n formulas (5.35) are introduced with the paper's own words, "We computed the three functions ... for n up to 20 and find that the following general expressions hold" (§5.3). This is an empirical induction/verification gap, not a circular reduction: it does not assume the target result, and the paper explicitly concedes in §7 that the divisor-sum coefficients "enter as boundary data fixed by matching onto the quasi-modular expansion, rather than being derived from the differential equation itself." That limitation weakens the abstract's "closed form at each order" claim but does not make the resummation equal to its inputs by construction. The exact differential equation (6.3) is derived from Bessel's equation and is then used as a consistency check; integration constants are fixed by matching the known small-y series, not by assuming the final resummation. The leading NP coefficient reproduces the independent result of [23], the Schwarzian partition function matches [28], the high-temperature expansion matches [19], and the saddle spectrum comparison uses the independent results of [25]. The only self-citations ([17], [26]) are incidental remarks and are not load-bearing. Overall, no specific equation reduces to its own input; the main risk is the unproven n-general pattern (5.35), which is a correctness/verification concern, not circularity.

Axiom & Free-Parameter Ledger

0 free parameters · 5 axioms · 0 invented entities

The paper's contribution is a reorganization of exact data, so its ledger is light: no free parameters, no invented entities, five named premises. No fitted constants: the combination y = 24π²/(λβ) is a kinematic scaling variable, and integration constants (e.g., C = −3 in (6.17)) are fixed by agreement with the exact expansion's leading coefficients — boundary data, not fits. Of the premises, the only ad hoc one is the induction that the pattern (5.35) extends beyond n = 20 — the load-bearing conjecture behind (5.38). The asymptotic-series/Borel premise is flagged by the authors themselves. The remaining premises are standard mathematics or prior exact results [5,24,25]. This is an honest ledger for a formal-theory paper; the risk is concentrated in the one conjectural premise, not in hidden fitting.

axioms (5)
  • domain assumption Exactness of the Bessel-function representation F(x) = (2/x)Σ(−1)ⁿ(2n+1)I_{2n+1}(x)q^{n(n+1)/2} and of the chord-space spectral representation (2.5).
    Quoted from prior literature [5,24] and used as the starting point of every derivation (§2, §4.2). Any normalization or double-scaling-limit subtlety in these representations would propagate through (4.22), (5.38), (6.3).
  • ad hoc to paper The general-n non-perturbative closed forms (5.35): f_{n,0} = −(3σ₁(n)/n)e^{ny/3}, plus f_{n,1} and f_{n,2} involving σ₃(n), hold for all n.
    Stated as "we computed ... for n up to 20 and find that the following general expressions hold" (§5.3). Only n = 1 is rigorously proven (Appendix A). Load-bearing for the resummation (5.38) and for the abstract's "closed form at each order in q̃" claim.
  • ad hoc to paper The large-x expansion (4.22), an asymptotic series with zero radius of convergence, can be reorganized via S-duality into the complete small-λ non-perturbative structure, with Borel-type corrections ~e^{−2x} negligible.
    The paper flags this itself: "The series (4.22) is asymptotic with zero radius of convergence ... This refinement is not studied further here" (§4.2 Remark). e^{−2x} = e^{−4β/√(λ(1−q))} is subdominant to q̃ⁿ in the λ→0, fixed-y regime, so the leading exponent spectrum is plausibly unaffected, but the "to all orders in λ" exactness is not quantized against it.
  • domain assumption On-shell action differences I_k − I_0 = 2π²k(k+1)/λ(1−2/β) of the bilocal-Liouville winding saddles, taken from [25] (their eq. 3.33a).
    Used in §7 for the exponent match. The match is exact on exponents; the saddle actions themselves are prior literature, and one-loop weights are not compared.
  • standard math Standard modular facts: SL(2,Z) transformation laws (4.11)–(4.15), Ramanujan identities (4.5), Jacobi triple product (4.18), Eisenstein q-series (4.2).
    Classical, quotable results used to derive (4.22) and (5.3)–(5.4); no assumptions beyond the well-known theorems.

pith-pipeline@v1.3.0-alltime-deepseek · 21051 in / 35909 out tokens · 298873 ms · 2026-08-02T06:46:38.340283+00:00 · methodology

0 comments
read the original abstract

We study the low-temperature expansion of the disk partition function $Z(\beta)$ of the double-scaled SYK model (DSSYK) at fixed coupling $\lambda=2p^{2}/N$, where $N$ is the number of Majorana fermions and $p$ is the number of fermions in each interaction term, both taken to infinity. We show that the exact Bessel-function representation of $Z(\beta)$, expanded at large argument (corresponding to low temperature), can be organized in terms of the classical ring of quasi-modular Eisenstein series $E_{2},E_{4},E_{6}$ and their differential identities. Exploiting the modular $S$-duality properties of this ring, we derive the semiclassical (small $\lambda$) low-temperature expansion of $Z(\beta)$, splitting it into a perturbative tower and a non-perturbative sector controlled by $\widetilde q=e^{-4\pi^{2}/\lambda}$. At each order in $\widetilde q$, we determine the non-perturbative correction in closed form up to second order in $\lambda$; the resulting series resums into a compact expression in the same Eisenstein series, extending previous semiclassical results beyond their strict $\beta\to\infty$ limit. We further show that this entire structure follows from a single, exact differential equation coupling a modular derivative to derivatives with respect to temperature. Finally, we prove that the non-perturbative sector of $Z(\beta)$ is exactly supported, to all orders in $\lambda$, on the same exponents as the on-shell actions of known bilocal-Liouville saddles of the DSSYK Schwarzian limit, pointing to a well-defined bulk origin for these non-perturbative corrections.

Review history (2 revisions) →

discussion (0)

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

Reference graph

Works this paper leans on

28 extracted references · 25 linked inside Pith

  1. [1]

    Sachdev and J

    S. Sachdev and J. Ye,Gapless Spin Fluid Ground State in a Random, Quantum Heisenberg Magnet,Phys. Rev. Lett.70(1993) 3339 [cond-mat/9212030]. S. Sachdev,Holographic Metals and the Fractionalized Fermi Liquid,Phys. Rev. Lett.105 (2010) 151602 [1006.3794]. 21

    Pith/arXiv arXiv 1993

  2. [2]

    Maldacena and D

    J. Maldacena and D. Stanford,Remarks on the Sachdev-Ye-Kitaev Model,Phys. Rev. D94 (2016) 106002 [1604.07818]

    Pith/arXiv arXiv 2016

  3. [3]

    Maldacena, D

    J. Maldacena, D. Stanford and Z. Yang,Conformal Symmetry and Its Breaking in Two Dimensional Nearly Anti-De-Sitter Space,PTEP2016(2016) 12C104 [1606.01857]

    Pith/arXiv arXiv 2016

  4. [4]

    Jensen,Chaos in AdS2 Holography,Phys

    K. Jensen,Chaos in AdS2 Holography,Phys. Rev. Lett.117(2016) 111601 [1605.06098]. G. Sárosi,AdS 2 holography and the SYK model,PoSModave2017(2018) 001 [1711.08482]

    Pith/arXiv arXiv 2016

  5. [5]

    Berkooz, M

    M. Berkooz, M. Isachenkov, V. Narovlansky and G. Torrents,Towards a Full Solution of the LargeNDouble-Scaled Syk Model,JHEP03(2019) 079 [1811.02584]

    Pith/arXiv arXiv 2019

  6. [6]

    T. G. Mertens, G. J. Turiaci and H. L. Verlinde,Solving the Schwarzian via the Conformal Bootstrap,JHEP08(2017) 136 [1705.08408]

    Pith/arXiv arXiv 2017

  7. [7]

    Engelsöy, T

    J. Engelsöy, T. G. Mertens and H. Verlinde,An investigation of AdS2 backreaction and holography,JHEP07(2016) 139 [1606.03438]. T. G. Mertens and G. J. Turiaci,Solvable models of quantum black holes: a review on Jackiw–Teitelboim gravity,Living Rev. Rel.26(2023) 4 [2210.10846]

    Pith/arXiv arXiv 2016

  8. [8]

    Erdős and D

    L. Erdős and D. Schröder,Phase Transition in the Density of States of Quantum Spin Glasses,Math. Phys. Anal. Geom.17(2014) 441 [1407.1552]. J. S. Cotler, G. Gur-Ari, M. Hanada, J. Polchinski, P. Saad, S. H. Shenker et al.,Black Holes and Random Matrices,JHEP05(2017) 118 [1611.04650]

    Pith/arXiv arXiv 2014

  9. [9]

    Berkooz, P

    M. Berkooz, P. Narayan and J. Simon,Chord Diagrams, Exact Correlators in Spin Glasses and Black Hole Bulk Reconstruction,JHEP08(2018) 192 [1806.04380]

    Pith/arXiv arXiv 2018

  10. [10]

    H. W. Lin,The Bulk Hilbert Space of Double Scaled SYK,JHEP11(2022) 060 [2208.07032]

    Pith/arXiv arXiv 2022

  11. [11]

    H. W. Lin and D. Stanford,A Symmetry Algebra in Double-Scaled Syk,SciPost Phys.15 (2023) 234 [2307.15725]

    Pith/arXiv arXiv 2023

  12. [12]

    Blommaert, T

    A. Blommaert, T. G. Mertens and S. Yao,Dynamical Actions and Q-Representation Theory for Double-Scaled Syk,JHEP02(2024) 067 [2306.00941]. A. Blommaert, T. G. Mertens and J. Papalini,The Dilaton Gravity Hologram of Double-Scaled Syk,JHEP06(2025) 050 [2404.03535]

    Pith/arXiv arXiv 2024

  13. [13]

    Blommaert, A

    A. Blommaert, A. Levine, T. G. Mertens, J. Papalini and K. Parmentier,An Entropic Puzzle in Periodic Dilaton Gravity and Dssyk,JHEP07(2025) 093 [2411.16922]

    Pith/arXiv arXiv 2025

  14. [14]

    M. P. Heller, J. Papalini and T. Schuhmann,Krylov Spread Complexity as Holographic Complexity Beyond Jackiw-Teitelboim Gravity,Phys. Rev. Lett.135(2025) 151602 [2412.17785]

    arXiv 2025

  15. [15]

    M. P. Heller, F. Ori, J. Papalini, T. Schuhmann and M.-T. Wang,De Sitter Holographic Complexity from Krylov Complexity in Dssyk,2510.13986

    arXiv

  16. [16]

    Bossi, L

    L. Bossi, L. Griguolo, J. Papalini, L. Russo and D. Seminara,Sine-Dilaton Gravity Vs Double-Scaled Syk: Exploring One-Loop Quantum Corrections,JHEP06(2025) 152 [2411.15957]. 22

    Pith/arXiv arXiv 2025

  17. [17]

    Alfinito and M

    E. Alfinito and M. Beccaria,Higher-Loop Wormhole Length in Sine-Dilaton Gravity from Dssyk Krylov Complexity,2606.20220

    Pith/arXiv arXiv

  18. [18]

    Fu, H.-S

    Y. Fu, H.-S. Jeong, K.-Y. Kim and J. F. Pedraza,Toward Krylov-Based Holography in Double-Scaled Syk,JHEP05(2026) 056 [2510.22658]

    Pith/arXiv arXiv 2026

  19. [19]

    Okuyama,High Temperature Expansion of Double Scaled Syk,Phys

    K. Okuyama,High Temperature Expansion of Double Scaled Syk,Phys. Lett. B843(2023) 138036 [2304.01522]

    Pith/arXiv arXiv 2023

  20. [20]

    M. J. Vergès,Cumulants of the q-semicircular law, tutte polynomials, and heaps,Canadian Journal of Mathematics65(2013) 863–878

    2013

  21. [21]

    A. Goel, V. Narovlansky and H. Verlinde,Semiclassical Geometry in Double-Scaled Syk, JHEP11(2023) 093 [2301.05732]

    Pith/arXiv arXiv 2023

  22. [22]

    Okuyama and K

    K. Okuyama and K. Suzuki,Correlators of Double Scaled Syk at One-Loop,JHEP05(2023) 117 [2303.07552]

    Pith/arXiv arXiv 2023

  23. [23]

    Okuyama,Non-Perturbative Corrections in the Semi-Classical Limit of Double-Scaled Syk, JHEP06(2025) 044 [2501.15501]

    K. Okuyama,Non-Perturbative Corrections in the Semi-Classical Limit of Double-Scaled Syk, JHEP06(2025) 044 [2501.15501]

    Pith/arXiv arXiv 2025

  24. [24]

    Xu,On Chord Dynamics and Complexity Growth in Double-Scaled SYK,JHEP06(2025) 259 [2411.04251]

    J. Xu,On Chord Dynamics and Complexity Growth in Double-Scaled SYK,JHEP06(2025) 259 [2411.04251]

    Pith/arXiv arXiv 2025

  25. [25]

    Berkooz, R

    M. Berkooz, R. Frumkin, O. Mamroud and J. Seitz,Twisted Times, the Schwarzian and Its Deformations in Dssyk,JHEP05(2025) 080 [2412.14238]

    Pith/arXiv arXiv 2025

  26. [26]

    Beccaria and A

    M. Beccaria and A. Cabo-Bizet,Giant Graviton Expansion of Schur Index and Quasimodular Forms,2403.06509

    Pith/arXiv arXiv

  27. [27]

    Billo, M

    M. Billo, M. Frau, L. Gallot, A. Lerda and I. Pesando,Modular anomaly equation, heat kernel and S-duality inN“2theories,JHEP11(2013) 123 [1307.6648]. B. Bajc and K. Trailović,Holographic thermal propagator from modularity,JHEP11(2025) 133 [2509.02226]

    Pith/arXiv arXiv 2013

  28. [28]

    Stanford and E

    D. Stanford and E. Witten,Fermionic Localization of the Schwarzian Theory,JHEP10 (2017) 008 [1703.04612]. 23

    Pith/arXiv arXiv 2017