pith. sign in

arxiv: 2605.28727 · v1 · pith:7SATBYZJnew · submitted 2026-05-27 · ⚛️ nucl-th · astro-ph.HE

Impact of hyperon mixing on neutron star structure based on Skyrme-type equations of state: Systematic analysis of Λ NN and ΛΛ N three-body forces with Bayesisan inference

Taeho Lee , Yoonhak Nam , Kazuyuki Sekizawa This is my paper

Pith reviewed 2026-06-29 09:20 UTC · model grok-4.3

classification ⚛️ nucl-th astro-ph.HE
keywords hyperonsneutron starsthree-body forcesSkyrme energy density functionalBayesian inferencemaximum massequation of state
0
0 comments X

The pith

Bayesian inference on Skyrme hyperon models favors sizable three-body repulsion to support observed neutron-star maximum masses.

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

The paper maps how density-dependent LambdaNN and LambdaLambdaN three-body terms in a Skyrme energy-density functional alter the onset density of hyperons and the stiffness of the post-onset equation of state in beta-equilibrated npe mu Lambda matter. Tolman-Oppenheimer-Volkoff integrations classify the resulting pressure-energy branches by monotonicity and number of extrema, while an exploratory Bayesian analysis with neutron-star mass-radius data alone assigns higher posterior weight to large repulsive values of the A3 and C3 parameters. SHAP diagnostics applied to an XGBoost surrogate confirm that these two parameters are the dominant controls on maximum mass and radius at 2 solar masses. The central result is that maximum-mass recovery is not achieved by a single mechanism; onset shifts, branch admissibility, and extremum structure must be examined together with the posterior distributions.

Core claim

Within the adopted likelihood constructed from neutron-star mass-radius observations and the chosen prior ranges, the posterior distributions favor sizable hyperonic three-body repulsion. The LambdaNN term simultaneously shifts the Lambda onset density and modifies the post-onset branch, while the LambdaLambdaN term leaves the onset unchanged but stiffens the finite-Lambda equation of state; increasing C3 raises M_max in admissible regions whereas increasing gamma reduces the stiffening at fixed C3. Representative two-extrema branches are connected by Maxwell constructions, and SHAP analysis identifies A3 and C3 as the leading controls of M_max and R_2.0.

What carries the argument

Density-dependent effective LambdaNN (beta, A3) and LambdaLambdaN (gamma, C3) three-body terms inside a Skyrme energy-density functional, inserted into beta-equilibrated npe mu Lambda matter and integrated via the Tolman-Oppenheimer-Volkoff equation, with posterior sampling and SHAP surrogate diagnostics.

If this is right

  • Increasing C3 at fixed gamma stiffens the post-onset branch and raises the maximum mass in mechanically stable regions.
  • The LambdaNN term shifts the Lambda onset density to lower values while simultaneously altering the post-onset stiffness.
  • For some reference interactions the parameter plane organizes into branch-limited regions and Maxwell-construction candidates.
  • SHAP values rank A3 and C3 as the dominant controls on both M_max and R_2.0 within the sampled posterior.

Where Pith is reading between the lines

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

  • If the favored repulsive three-body terms prove incompatible with hypernuclear binding energies, additional density-dependent or many-body terms will be required.
  • The same parameter plane could be re-weighted with tidal-deformability or cooling data to test whether the posterior remains concentrated at large A3 and C3.
  • The classification of branches by extremum count supplies a diagnostic that could be applied to other hyperon models to separate onset-driven from stiffness-driven mass recovery.

Load-bearing premise

The Skyrme energy-density functional with only npe mu Lambda composition and a likelihood built solely from neutron-star mass-radius data is assumed to capture the essential physics without additional nuclear or astrophysical constraints.

What would settle it

A precise measurement showing that the radius at 2 solar masses lies outside the narrow band produced by the high-posterior A3-C3 region, or the discovery of a neutron star whose mass exceeds the maximum allowed by any admissible branch once the favored repulsive parameters are fixed.

Figures

Figures reproduced from arXiv: 2605.28727 by Kazuyuki Sekizawa, Taeho Lee, Yoonhak Nam.

Figure 1
Figure 1. Figure 1: FIG. 1. Representative EOS branches in the [PITH_FULL_IMAGE:figures/full_fig_p008_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2. Representative EOS bands showing the response to the [PITH_FULL_IMAGE:figures/full_fig_p011_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: (b) shows that many of these extrema are not encoun￾tered by the neutron-star branch before the maximum mass is determined. These regions should therefore not be interpreted as instabilities of the neutron-star configurations that deter￾mine Mmax. A smaller part of the SΛΛ1 ′ parameter plane has N NS ext = 1, as shown in [PITH_FULL_IMAGE:figures/full_fig_p012_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: FIG. 4. Representative EOS bands showing the response to the [PITH_FULL_IMAGE:figures/full_fig_p014_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5. Maps of the [PITH_FULL_IMAGE:figures/full_fig_p015_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: FIG. 6. Representative Maxwell construction and the corresponding [PITH_FULL_IMAGE:figures/full_fig_p016_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: FIG. 7. Marginalized one-dimensional posterior distributions for the seven-dimensional [PITH_FULL_IMAGE:figures/full_fig_p019_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: FIG. 8. Marginalized one-dimensional posterior distributions for the six-dimensional [PITH_FULL_IMAGE:figures/full_fig_p019_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: FIG. 9. Weighted posterior mass–radius curves for the four ex [PITH_FULL_IMAGE:figures/full_fig_p020_9.png] view at source ↗
Figure 11
Figure 11. Figure 11: FIG. 11. SHAP beeswarm plots for the XGBoost surrogate model [PITH_FULL_IMAGE:figures/full_fig_p021_11.png] view at source ↗
read the original abstract

We study hyperonic density-dependent three-body effects in cold neutron-star matter using a Skyrme energy-density-functional framework. In beta-equilibrated $npe\mu\Lambda$ matter, the effective $\Lambda NN$ and $\Lambda\Lambda N$ terms are varied separately in the $(\beta,A_3)$ and $(\gamma,C_3)$ planes, and each tabulated equation of state is used in Tolman--Oppenheimer--Volkoff calculations. The calculated $P$--$\varepsilon$ branches are classified by monotonicity and extremum structure. The $\Lambda\Lambda N$ term does not affect the $\Lambda$-onset condition, but modifies the finite-$\Lambda$ post-onset EOS: increasing $C_3$ generally stiffens the post-onset branch and raises $M_{\max}$ in mechanically admissible regions, whereas increasing $\gamma$ reduces this enhancement at fixed $C_3$. In contrast, the $\Lambda NN$ term shifts the $\Lambda$-onset density and modifies the post-onset EOS simultaneously, producing organized branch-limited and Maxwell-candidate regions for some reference interactions. Representative two-extrema cases are examined with Maxwell constructions. We also perform an exploratory Bayesian analysis using neutron-star mass--radius information alone and apply XGBoost--SHAP surrogate diagnostics to summarize parameter sensitivities. Within the adopted likelihood and prior ranges, the posterior weight tends to favor sizable hyperonic three-body repulsion, and the SHAP analysis identifies $A_3$ and $C_3$ as important controls of $M_\text{max}$ and $R_{2.0}$. These results show that maximum-mass recovery in hyperonic neutron stars is not a single mechanism: $M_\text{max}$ maps must be interpreted together with onset behavior, branch admissibility, and extremum-count diagnostics. *shortened due to the arXiv's word limit.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper studies hyperonic three-body forces (ΛNN and ΛΛN) in Skyrme EDFs for cold beta-equilibrated npeμΛ neutron-star matter. It systematically varies the (β, A3) and (γ, C3) parameters, tabulates the resulting EOS, solves the TOV equation, classifies P–ε branches by monotonicity and extremum structure, performs representative Maxwell constructions, and conducts an exploratory Bayesian analysis using only neutron-star mass–radius data together with XGBoost–SHAP diagnostics. The central claim is that, within the adopted priors and M–R likelihood, the posterior favors sizable hyperonic three-body repulsion and that A3 and C3 are the dominant controls of M_max and R_2.0.

Significance. If the results hold, the work demonstrates that maximum-mass recovery in hyperonic stars is not a single-parameter mechanism but requires joint consideration of onset density, post-onset stiffness, and branch admissibility. The systematic mapping of the two parameter planes and the use of SHAP to rank sensitivities constitute useful diagnostics for EDF-based hyperonic models. The explicit branch-classification procedure is a concrete methodological contribution.

major comments (2)
  1. [Bayesian analysis section] Bayesian analysis section (likelihood construction): The likelihood is assembled exclusively from neutron-star mass–radius constraints. No term enforces the empirical Λ single-particle potential depth (≈ −30 MeV at saturation) or ΛN scattering lengths. Because A3 and C3 directly set both the Λ onset density and the post-onset repulsion, the reported posterior preference for sizable repulsion and the SHAP ranking of A3, C3 may be driven by the chosen prior volume and the limited M–R information rather than by nuclear data.
  2. [Parameter variation and SHAP diagnostics] Parameter variation and SHAP diagnostics: The same parameters (β, A3, γ, C3) that define the EOS are also the objects of Bayesian inference; the SHAP surrogate is trained on quantities derived from those same EOS. This creates a moderate circularity that must be quantified (e.g., by reporting the effective prior volume after the branch-admissibility filter) before the claim that the posterior “tends to favor sizable hyperonic three-body repulsion” can be regarded as robust.
minor comments (2)
  1. [Abstract] Abstract: “Bayesisan” is a typographical error.
  2. [EOS construction] The manuscript states that the ΛΛN term does not affect the Λ-onset condition; an explicit equation or numerical check confirming this independence would strengthen the presentation.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive and detailed report. The two major comments concern the scope of the Bayesian analysis and the interpretation of the SHAP diagnostics. We address each point below and indicate where revisions will be made to improve clarity.

read point-by-point responses

  1. Referee: [Bayesian analysis section] Bayesian analysis section (likelihood construction): The likelihood is assembled exclusively from neutron-star mass–radius constraints. No term enforces the empirical Λ single-particle potential depth (≈ −30 MeV at saturation) or ΛN scattering lengths. Because A3 and C3 directly set both the Λ onset density and the post-onset repulsion, the reported posterior preference for sizable repulsion and the SHAP ranking of A3, C3 may be driven by the chosen prior volume and the limited M–R information rather than by nuclear data.

    Authors: We agree that the Bayesian section uses only neutron-star mass–radius constraints and does not include nuclear-physics anchors such as the empirical Λ single-particle potential or scattering lengths. The manuscript already describes the analysis as exploratory and qualifies the posterior preference as holding “within the adopted likelihood and prior ranges.” To make this limitation explicit, we will add a short paragraph in the Bayesian section stating that the reported trends are conditioned solely on the M–R data and chosen priors, and that inclusion of nuclear constraints is left for future work. This is a clarification rather than a change in methodology or conclusions. revision: partial

  2. Referee: [Parameter variation and SHAP diagnostics] Parameter variation and SHAP diagnostics: The same parameters (β, A3, γ, C3) that define the EOS are also the objects of Bayesian inference; the SHAP surrogate is trained on quantities derived from those same EOS. This creates a moderate circularity that must be quantified (e.g., by reporting the effective prior volume after the branch-admissibility filter) before the claim that the posterior “tends to favor sizable hyperonic three-body repulsion” can be regarded as robust.

    Authors: The SHAP analysis is applied after the posterior has been obtained; it ranks the influence of the input parameters on the derived observables (M_max, R_2.0) and is a standard interpretability tool rather than a source of circularity. Nevertheless, we accept the request to quantify the effective prior volume after the branch-admissibility filter. In the revised manuscript we will report the retained fraction of the prior volume (approximately 35–40 % of the sampled points survive the monotonicity and extremum criteria) so that readers can assess how strongly the admissibility cuts shape the posterior. This addition addresses the robustness concern directly. revision: partial

Circularity Check

0 steps flagged

No significant circularity; Bayesian update uses external M-R likelihood

full rationale

The paper varies Skyrme parameters (β, A3, γ, C3) to build EOS tables, solves TOV equations, classifies branches, then constructs a likelihood solely from independent neutron-star mass-radius observations to obtain the posterior. SHAP is applied post-inference as a sensitivity diagnostic on the sampled parameters versus derived outputs (M_max, R_2.0). No equation reduces a claimed result to its own inputs by construction, no fitted quantity is relabeled as a prediction, and no self-citation chain is invoked as load-bearing justification. The derivation chain is therefore self-contained against the stated external data.

Axiom & Free-Parameter Ledger

2 free parameters · 2 axioms · 0 invented entities

The central results rest on the Skyrme EDF being a valid description of hyperonic matter and on the chosen Bayesian likelihood being representative; the four parameters β, A3, γ, C3 are treated as free and scanned without external calibration shown in the abstract.

free parameters (2)
  • β, A3
    Varied independently in the (β,A3) plane to control the ΛNN three-body term; directly affect onset density and post-onset stiffness.
  • γ, C3
    Varied independently in the (γ,C3) plane to control the ΛΛN three-body term; affect post-onset EOS only.
axioms (2)
  • domain assumption Skyrme energy-density-functional framework is adequate for cold beta-equilibrated npeμΛ matter
    Used as the base model for all tabulated EOS branches.
  • domain assumption Neutron-star mass-radius data alone provide a sufficient likelihood for constraining the three-body parameters
    Explicitly stated as the basis for the exploratory Bayesian analysis.

pith-pipeline@v0.9.1-grok · 5903 in / 1562 out tokens · 36331 ms · 2026-06-29T09:20:59.720936+00:00 · methodology

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

37 extracted references · 11 canonical work pages · 6 internal anchors

  1. [1]

    A. Gal, E. V . Hungerford, and D. J. Millener,Strangeness in nu- clear physics, Reviews of Modern Physics88, 035004 (2016), arXiv:1605.00557 [nucl-th]

    work page internal anchor Pith review Pith/arXiv arXiv 2016

  2. [2]

    Tolos and L

    L. Tolos and L. Fabbietti,Strangeness in nuclei and neutron stars, Progress in Particle and Nuclear Physics112, 103770 (2020), arXiv:2002.09223 [nucl-ex]

    work page arXiv 2020

  3. [3]

    P. B. Demorest, T. Pennucci, S. M. Ransom, M. S. E. Roberts, and J. W. T. Hessels,A two-solar-mass neutron star measured using shapiro delay, Nature467, 1081–1083 (2010)

    2010

  4. [4]

    Antoniadis, P

    J. Antoniadis, P. C. C. Freire, N. Wex, T. M. Tauris, R. S. Lynch, M. H. van Kerkwijk, M. Kramer, C. Bassa, V . S. Dhillon, T. Driebe, J. W. T. Hessels, V . M. Kaspi, V . I. Kondratiev, N. Langer, T. R. Marsh, M. A. McLaughlin, T. T. Pennucci, S. M. Ransom, I. H. Stairs, J. van Leeuwen, J. P. W. Verbiest, and D. G. Whelan,A massive pulsar in a compact rel...

    2013

  5. [5]

    Fonseca, H

    E. Fonseca, H. T. Cromartie, T. T. Pennucci, P. S. Ray, A. Y . Kirichenko, S. M. Ransom, P. B. Demorest, I. H. Stairs, Z. Ar- zoumanian, L. Guillemot,et al.,Refined mass and geometric measurements of the high-mass PSR J0740+6620, The Astro- physical Journal Letters915, L12 (2021), arXiv:2104.00880 [astro-ph.HE]

    work page arXiv 2021

  6. [6]

    T. E. Riley, A. L. Watts, S. Bogdanov, P. S. Ray, R. M. Lud- lam, S. Guillot, Z. Arzoumanian, C. L. Baker, A. V . Bilous, D. Chakrabarty, K. C. Gendreau, A. K. Harding, W. C. G. Ho, J. M. Lattimer, S. M. Morsink, and T. E. Strohmayer,A NICER view of PSR J0030+0451: Millisecond pulsar parameter esti- mation, The Astrophysical Journal Letters887, L21 (2019)

    2019

  7. [7]

    T. E. Riley, A. L. Watts, P. S. Ray, S. Bogdanov, S. Guillot, S. M. Morsink, A. V . Bilous, Z. Arzoumanian, D. Choudhury, J. S. Deneva, K. C. Gendreau, A. K. Harding, W. C. G. Ho, J. M. Lattimer, M. Loewenstein, R. M. Ludlam, C. B. Markwardt, T. Okajima, C. Prescod-Weinstein, R. A. Remillard, M. T. Wolff, E. Fonseca, H. T. Cromartie, M. Kerr, T. T. Pennuc...

    2021

  8. [8]

    Choudhury, T

    D. Choudhury, T. H. J. Salmi, S. Vinciguerra, T. E. Riley, Y . Kini, A. L. Watts, B. Dorsman, S. Bogdanov, S. Guil- lot, P. S. Ray, D. J. Reardon, R. A. Remillard, A. V . Bilous, D. Huppenkothen, J. M. Lattimer, N. Rutherford, Z. Arzou- manian, K. C. Gendreau, S. M. Morsink, and W. C. G. Ho,A NICER view of the nearest and brightest millisecond pulsar: PSR...

    work page arXiv 2024

  9. [9]

    B. P. Abbottet al.,Gw170817: Measurements of neutron star radii and equation of state, Physical Review Letters121, 161101 (2018)

    2018

  10. [10]

    An Effective Equation of State for Dense Matter with Strangeness

    S. Balberg and A. Gal,An effective equation of state for dense matter with strangeness, Nuclear Physics A625, 435–472 (1997), arXiv:nucl-th/9704013

    work page internal anchor Pith review Pith/arXiv arXiv 1997

  11. [11]

    Do hyperons exist in the interior of neutron stars ?

    D. Chatterjee and I. Vidaña,Do hyperons exist in the interior of neutron stars?, The European Physical Journal A52, 29 (2016), arXiv:1510.06306 [nucl-th]

    work page internal anchor Pith review Pith/arXiv arXiv 2016

  12. [12]

    Schulze and E

    H.-J. Schulze and E. Hiyama,Skyrme force for light and heavy hypernuclei, Physical Review C90, 047301 (2014)

    2014

  13. [13]

    Guleria, S

    N. Guleria, S. K. Dhiman, and R. Shyam,A study ofΛhyper- nuclei within the skyrme–hartree–fock model, Nuclear Physics A886, 71–91 (2012)

    2012

  14. [14]

    D. E. Lanskoy,Double-Λhypernuclei in the skyrme-hartree- fock approach and nuclear core polarization, Phys. Rev. C58, 3351–3358 (1998)

    1998

  15. [15]

    Minato and S

    F. Minato and S. Chiba,Fission barrier of actinide nuclei with double-Λparticles within the skyrme–hartree–fock method, Nuclear Physics A856, 55–67 (2011)

    2011

  16. [16]

    Jinno, K

    A. Jinno, K. Murase, Y . Nara, and A. Ohnishi,RepulsiveΛ potentials in dense neutron star matter and binding energy ofΛin hypernuclei, Physical Review C108, 065803 (2023), arXiv:2306.17452 [nucl-th]

    work page arXiv 2023

  17. [17]

    Malik and C

    T. Malik and C. Providência,Bayesian inference of signa- tures of hyperons inside neutron stars, Physical Review D106, 063024 (2022)

    2022

  18. [18]

    Huang, L

    C. Huang, L. Tolos, C. Providência, and A. L. Watts,Constrain- ing a relativistic mean field model using neutron star mass- radius measurements ii: Hyperonic models, Monthly Notices of the Royal Astronomical Society536, 3262–3275 (2025)

    2025

  19. [19]

    X. D. Sun, S. C. Han, J. N. Hu, and A. Li,Exploring hyperon Skyrme forces in multi-Λhypernuclei and neutron star mat- ter, Monthly Notices of the Royal Astronomical Society546, stag233 (2026), arXiv:2602.03388 [nucl-th]

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  20. [20]

    Tanimura, C

    Y . Tanimura, C. Hyun, and M. Cheoun, Constraining the ΛΛinteraction with terrestrial and astronomical data (2026), arXiv:2602.18356 [nucl-th]

    work page arXiv 2026

  21. [21]

    T. H. R. Skyrme,The effective nuclear potential, Nuclear Physics9, 615–634 (1958)

    1958

  22. [22]

    Vautherin and D

    D. Vautherin and D. M. Brink,Hartree-fock calculations with skyrme’s interaction. i. spherical nuclei, Physical Review C5, 626–647 (1972)

    1972

  23. [23]

    Rayet,Self-consistent calculations of hypernuclear proper- ties with the Skyrme interaction, Annals of Physics102, 226– 251 (1976)

    M. Rayet,Self-consistent calculations of hypernuclear proper- ties with the Skyrme interaction, Annals of Physics102, 226– 251 (1976)

    1976

  24. [24]

    Rayet,Skyrme parametrization of an effectiveΛ-nucleon in- teraction, Nuclear Physics A367, 381–397 (1981)

    M. Rayet,Skyrme parametrization of an effectiveΛ-nucleon in- teraction, Nuclear Physics A367, 381–397 (1981)

    1981

  25. [25]

    Bender, P.-H

    M. Bender, P.-H. Heenen, and P.-G. Reinhard,Self-consistent mean-field models for nuclear structure, Reviews of Modern Physics75, 121–180 (2003)

    2003

  26. [26]

    Schulze, inThe 13th International Conference on Hyper- Nuclear and Strange Particle Physics: HYP2018, AIP Confer- ence Proceedings, V ol

    H.-J. Schulze, inThe 13th International Conference on Hyper- Nuclear and Strange Particle Physics: HYP2018, AIP Confer- ence Proceedings, V ol. 2130 (AIP Publishing, 2019) p. 020009

    2019

  27. [27]

    Yamamoto, H

    Y . Yamamoto, H. Band¯o, and J. Žofka,On theΛ-hypernuclear single particle energies, Progress of Theoretical Physics80, 757–761 (1988)

    1988

  28. [28]

    N. K. Glendenning,Compact Stars: Nuclear Physics, Particle Physics, and General Relativity, 2nd ed., Astronomy and As- trophysics Library (Springer, New York, 2000)

    2000

  29. [29]

    Equations of state for supernovae and compact stars

    M. Oertel, M. Hempel, T. Klähn, and S. Typel,Equations of state for supernovae and compact stars, Reviews of Modern Physics89, 015007 (2017), arXiv:1610.03361 [astro-ph.HE]

    work page internal anchor Pith review Pith/arXiv arXiv 2017

  30. [30]

    R. C. Tolman,Static solutions of Einstein’s field equations for spheres of fluid, Physical Review55, 364–373 (1939)

    1939

  31. [31]

    J. R. Oppenheimer and G. M. V olkoff,On massive neutron cores, Physical Review55, 374–381 (1939)

    1939

  32. [32]

    N. K. Glendenning,First-order phase transitions with more than one conserved charge: Consequences for neutron stars, Physical Review D46, 1274–1287 (1992)

    1992

  33. [33]

    T. E. Riley, G. Raaijmakers, and A. L. Watts,On parametrized cold dense matter equation-of-state inference, Monthly Notices of the Royal Astronomical Society478, 1093–1131 (2018)

    2018

  34. [34]

    Raaijmakers, N

    G. Raaijmakers, N. Rutherford, P. Timmerman, T. H. J. Salmi, A. L. Watts, C. Prescod-Weinstein, I. Svensson, and M. Mendes,NEoST: A python package for nested sampling of the neutron star equation of state, Journal of Open Source Soft- ware10, 6003 (2025)

    2025

  35. [35]

    J. S. Speagle,dynesty: a dynamic nested sampling package for estimating bayesian posteriors and evidences, Monthly Notices of the Royal Astronomical Society493, 3132–3158 (2020)

    2020

  36. [36]

    Chen and C

    T. Chen and C. Guestrin, inProceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining(2016) pp. 785–794

    2016

  37. [37]

    S. M. Lundberg and S.-I. Lee, inAdvances in Neural Informa- tion Processing Systems, V ol. 30 (2017) arXiv:1705.07874

    work page internal anchor Pith review Pith/arXiv arXiv 2017