Pith. sign in

REVIEW 2 major objections 5 minor 70 references

Hypergraph backboning

T0 review · 2 major / 5 minor · reviewed 2026-07-12 · grok-4.5

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

Pith's one-line read An information-theoretic method prunes nested and redundant hyperedges to extract a minimal higher-order backbone that preserves essential structure under local heterogeneity.

desk verdict Solid MDL hypergraph sparsifier that actually handles local order heterogeneity and weights; modeling choice is real but operationally checked. read the letter →

arxiv 2606.00893 v2 pith:RPCRCU6O submitted 2026-05-30 cs.SI physics.soc-ph

Alec Kirkley , Helcio Felippe , Federico Malizia , Federico Battiston This is my paper · ORCID
classification cs.SIphysics.soc-ph
keywords hypergraphbackboningminimumdescriptionlengthhigher-ordernetworkssparsificationnestedhyperedgesweightedhypergraphsreducedmutualinformation
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

Hypergraphs capture multi-body interactions that ordinary graphs miss, but they are high-dimensional and full of nested or overlapping redundancies. This paper shows how to treat sparsification as lossless compression: transmit a small set of parent hyperedges first, then encode the rest as children that share nodes with those parents. The best backbone is the subset that minimizes the total description length. Because selection is local rather than global, different interaction orders can be kept in different regions, so connectivity is not sacrificed for compression. The same objective extends to weighted hypergraphs by adding a simple weight prior that rewards strong interactions. On synthetic data the method recovers planted nested structure; on real systems it typically retains only one-quarter to one-third of the hyperedges while still capturing the core architecture.

What carries the argument

The description-length objective L(G,B) = L(B) + L(G|B), rewritten via reduced mutual information R(c,p) that increases with hyperedge overlap and is positive for nested pairs; greedy optimization over the intersection graph finds an approximate minimizer.

What would settle it

Construct a synthetic hypergraph whose scientifically essential hyperedges are deliberately non-overlapping and non-nested, then check whether the MDL backbone still recovers them or instead retains only the compressible but inessential ones.

Watch

Extended reading notes

Core claim

The minimum-description-length backbone B* = arg min L(G,B), built from combinatorial parent costs and conditional child costs that reward node overlap and nestedness, yields a sparse higher-order representation that recovers planted structure on controlled hypergraphs and substantially sparsifies empirical hypergraphs without discarding essential connectivity.

Load-bearing premise

The claim rests on treating one particular combinatorial encoding of sizes and overlaps as the true measure of structural redundancy, so that what compresses best is assumed to be what matters scientifically.

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

2 major / 5 minor

Summary. The paper introduces a minimum-description-length (MDL) framework for extracting a structural backbone B* from an unweighted or weighted hypergraph G. The objective L(G,B) (Eqs. 5–6) encodes parent hyperedges with fixed-length combinatorial costs H(p) and transmits remaining child hyperedges via conditional costs H(c|p) that exploit node overlap and nestedness; the associated reduced mutual information R(c,p) (Eq. 11) is shown to increase with overlap and to be positive for nested pairs under mild size conditions (Appendix A). The same construction extends to weighted hypergraphs by adding an empirical-Bayes weight prior controlled by a single ratio parameter γ (Eqs. 13–14). Greedy node- and edge-addition heuristics on the intersection graph are supplied (Appendix D). Controlled synthetic experiments recover planted top faces and parent–child structure under order heterogeneity and noise (Fig. 2), while application to 21 empirical hypergraphs yields consistent sparsification (roughly one-quarter to one-third of hyperedges retained, inverse compression η ≈ 0.54–0.83; Table I) together with more balanced layer-size distributions (Fig. 4).

Significance. If the central claim holds, the work supplies the first fully nonparametric, information-theoretic sparsifier that operates simultaneously across all interaction orders and that naturally incorporates edge weights. The combinatorial derivation is clean, the RMI monotonicity/positivity proofs are machine-checkable in spirit, synthetic recovery is strong with error bars, and the greedy optimizer matches exact enumeration on small samples (Fig. 6a). Code is released inside Hypergraphx and the empirical corpus is public, making the contribution immediately usable for higher-order network analysis. These strengths place the paper above the typical threshold for a methods contribution in network science.

major comments (2)
  1. Section II, Eqs. (1)–(4) and (11): the fixed-length combinatorial encoding of size, node sets and overlap is a modeling choice that defines what counts as “structural redundancy.” While Appendix A proves that the resulting RMI rewards nestedness and overlap, and while synthetic recovery is excellent, the paper never independently justifies why this particular code (rather than, e.g., a degree-corrected or community-aware code) preserves scientifically essential connectivity rather than merely compressible patterns. A short discussion or an ablation against an alternative encoding would strengthen the claim that B* retains “core structural information.”
  2. Appendix D and Fig. 6: the combinatorial argmin of L(G,B) is approximated by greedy star-partition construction on the intersection graph. Exact enumeration on random-walk samples of size ~15 shows near-optimality, yet no approximation guarantee or systematic gap analysis is given for the full empirical instances (some with |G| > 10^5). Because the backbone itself is defined as the MDL optimum, a clearer statement of the residual sub-optimality risk (or a more expensive Monte-Carlo check on a subset of the larger networks) is needed before the empirical sparsification ratios can be treated as definitive.
minor comments (5)
  1. Figure 1 caption and surrounding text: the visual distinction between the reducibility result R* and the proposed backbone B* is clear, but the color legend for hyperedge orders is not repeated in later synthetic figures; a single consistent palette would help.
  2. Equation (8) and Table I: the inverse compression ratio η is reported to two decimals; given that L is measured in bits under a fixed-length code, reporting one more significant figure (or the absolute bit savings) would make the magnitude of compression easier to interpret.
  3. Section III, Eqs. (22)–(25): the weight threshold w*(γ) is derived for both Poisson and geometric priors, yet the main empirical weighted experiments (Appendix F) only show retained-edge fractions. Adding the realized w* values for the chosen γ would let readers see how often weight alone drives inclusion.
  4. Appendix E, Table II: the local (node-neighborhood) variant systematically retains more edges and yields higher η; a one-sentence recommendation on when the global versus local procedure is preferable would be useful for practitioners.
  5. Minor typographical points: “po” / “pf” / “pn” noise parameters are introduced without a unified notation table; “invs13/15” dataset names appear without expansion; a few sentences in the introduction repeat the phrase “higher-order interactions” three times in close succession.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: MDL objective is constructed from first-principles combinatorial codes and validated by recovery of externally planted structure, not by definitional self-reference.

full rationale

The central derivation (Sec. II) builds L(G,B) = L(B) + L(G|B) from fixed-length combinatorial codes H(p) and H(c|p) that count possible sizes, node sets, and overlaps (Eqs. 1–4); B* is then defined as the argmin of this objective (Eq. 7). This is the ordinary MDL construction, not a reduction of a claimed prediction to its own inputs. The RMI rewrite (Eqs. 9–12) and Appendix A monotonicity/positivity proofs are internal algebraic consequences of the same codes, not circular imports. Synthetic experiments plant top faces or parent–child pairs independently of the objective and measure recovery via hypergraph NMI against those planted objects (Fig. 2), outperforming both the authors’ prior reducibility method and a simple top-layer baseline; empirical sparsification (Table I, Fig. 4) is reported as an observed outcome, not a forced prediction. Weighted extension introduces a single free hyperparameter γ whose effect is explored rather than fitted-and-predicted. Self-citations ([36], [27], [56]) supply comparison baselines or prior NMI definitions and are not load-bearing uniqueness theorems that force the present result. Consequently the derivation chain does not collapse by construction; the only residual self-reference is the definitional character of any MDL objective, which is expected and does not elevate the score above 1.

Assumptions & free parameters 2 free parameters · 5 assumptions · 2 invented entities

The central claim rests on the MDL principle applied to a particular combinatorial encoding of hyperedges and overlaps, plus a greedy star-partition optimizer and (for weights) an empirical-Bayes mean constraint with free parameter γ. These are modeling and algorithmic choices, not theorems forced by data. No new physical entities are postulated; the ‘backbone’ is an inferred subset defined by the objective.

free parameters (2)
  • γ (weight prior mean ratio) = user-chosen; examples 0.1, 0.5, 0.01 used in figures
    Tunable hyperparameter in (0,1] that sets the relative expected weight of child versus parent hyperedges; controls how strongly weight, versus topology, pushes an edge into the backbone. Required for the weighted objective.
  • weight prior family (Poisson vs geometric)
    Choice of distribution family for L(w(e), b_e); both are fully specified by their means, but the family is a modeling decision that affects the weight reward term.
assumptions (5)
  • domain assumption MDL / Kraft inequality: best backbone is the subset B minimizing total codelength L(B)+L(G|B) under a valid prefix code.
    Invoked throughout Section II as the optimality criterion; standard in information-theoretic model selection but not forced by the data.
  • ad hoc to paper Fixed-length combinatorial codes for hyperedge size (log L bits) and node sets (log binom(N,|e|) bits), and the specific conditional encoding of a child via overlap size and residual nodes (Eq. 3).
    Defines H(p) and H(c|p); alternative codes or generative models would yield different backbones. Introduced in Section II.
  • ad hoc to paper Parent-child relations form a star partition of the intersection graph (each child has exactly one parent; parents have no parents).
    Structural constraint used for both the objective and the greedy optimizers (Appendix D).
  • domain assumption Empirical-Bayes constraint that the global mean weight is reproduced by averaging parent and child means, yielding μ_be(γ, w̄).
    Section III; standard empirical-Bayes device but still a modeling choice that fixes the weight priors given γ.
  • ad hoc to paper Greedy node- or edge-addition on the intersection graph sufficiently approximates the combinatorial argmin of L(G,B).
    Appendix D; exact only for tiny samples; large-graph claims rest on this approximation.
invented entities (2)
  • Hyperedge reduced mutual information R(c,p) (Eq. 11) and its weighted extension R_w
    purpose: Quantifies information shared by a parent-child pair so that L can be rewritten as L(G,G) minus sum of RMI terms; drives which edges become parents.
    Constructed for this paper from contingency-table mutual information minus log min(|p|,|c|); no independent external measurement.
  • MDL hypergraph backbone B* (and local neighborhood variant B^(local))
    purpose: The inferred sparse subset claimed to retain essential higher-order structure.
    Defined as the minimizer of the paper’s own objective; validated only inside the paper’s synthetic and compression experiments.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Hypergraph backboning." pith.science (2026). https://pith.science/paper/RPCRCU6O

@misc{pith2026260600893,
  author       = {Pith},
  title        = {Pith review of: Hypergraph backboning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RPCRCU6O}},
  note         = {Machine review of arXiv:2606.00893}
}
read the original abstract

Hypergraphs provide a natural framework for describing complex networked systems with higher-order, non-dyadic interactions. Due to their high dimensionality and often redundant structure, a key challenge is to develop methods that simplify hypergraph representations while preserving the essential structure of interactions. Here we present a principled, efficient, and non-parametric information-theoretic method for pruning nested and/or redundant structures in hypergraphs, enabling a minimal representation of higher-order interactions in the presence of local heterogeneity. Our approach naturally extends to weighted hypergraphs, where higher-order topology and hyperedge weights combine to identify the system's structural backbone. We validate the method on controlled synthetic hypergraphs and apply it to empirical datasets from diverse domains, demonstrating substantial sparsification without loss of core structural information.

Figures

Figures reproduced from arXiv: 2606.00893 by the authors.

Figure 1
Figure 1. FIG. 1 [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2 [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3 [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: FIG. 4 [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: FIG. 5 [PITH_FULL_IMAGE:figures/full_fig_p014_5.png]
Figure 6
Figure 6. Figure 6: FIG. 6 [PITH_FULL_IMAGE:figures/full_fig_p017_6.png]
Figure 7
Figure 7. Figure 7: FIG. 7 [PITH_FULL_IMAGE:figures/full_fig_p018_7.png]
Figure 8
Figure 8. Figure 8: shows the results of applying the backboning method of Sec. III to these contact-based hypergraphs. For each dataset, we plot the fraction of hyperedges retained after applying the weighted backboning method, for different choices of the weight prior and values of γ. A…

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

70 extracted references · 3 linked inside Pith

  1. [1]

    Battiston, G

    F. Battiston, G. Cencetti, I. Iacopini, V. Latora, M. Lu- cas, A. Patania, J.-G. Young, and G. Petri, Networks beyond pairwise interactions: Structure and dynamics. Phys. Rep.874, 1–92 (2020)

    2020

  2. [2]

    Battiston, E

    F. Battiston, E. Amico, A. Barrat, G. Bianconi, G. Fer- raz de Arruda, B. Franceschiello, I. Iacopini, S. K´ efi, V. Latora, Y. Moreno, et al., The physics of higher- order interactions in complex systems.Nat. Phys.17(10), 1093–1098 (2021)

    2021

  3. [3]

    Bianconi, Higher-order networks

    G. Bianconi, Higher-order networks. Cambridge Univer- sity Press (2021)

    2021

  4. [4]

    C. Bick, E. Gross, H. A. Harrington, and M. T. Schaub, What are higher-order networks?SIAM Rev.65(3), 686– 731 (2023)

    2023

  5. [5]

    Berge, Hypergraphs: Combinatorics of Finite Sets

    C. Berge, Hypergraphs: Combinatorics of Finite Sets. Elsevier (1984)

    1984

  6. [6]

    A. R. Benson, R. Abebe, M. T. Schaub, A. Jadbabaie, and J. Kleinberg, Simplicial closure and higher-order link prediction.Proc. Natl. Acad. Sci. U.S.A.115(48), E11221–E11230 (2018)

    2018

  7. [7]

    Petri and A

    G. Petri and A. Barrat, Simplicial activity driven model. Phys. Rev. Lett.121, 228301 (2018)

    2018

  8. [8]

    Contisciani, F

    M. Contisciani, F. Battiston, and C. De Bacco, Infer- ence of hyperedges and overlapping communities in hy- pergraphs.Nat. Commun.13, 7228 (2022)

    2022

Show all 70 references
  1. [9]

    Di Gaetano, F

    L. Di Gaetano, F. Battiston, and M. Starnini, Percola- tion and topological properties of temporal higher-order networks.Phys. Rev. Lett.132, 037401 (2024)

    2024

  2. [10]

    Battiston, C

    F. Battiston, C. Bick, M. Lucas, A. P. Mill´ an, P. S. Skardal, and Y. Zhang, Collective dynamics on higher- order networks.Nat. Rev. Phy.8, 146 (2026)

    2026

  3. [11]

    Iacopini, G

    I. Iacopini, G. Petri, A. Barrat, and V. Latora, Simplicial models of social contagion.Nat. Commun.10(1), 2485 (2019)

    2019

  4. [12]

    Burgio, S

    G. Burgio, S. G´ omez, and A. Arenas, Triadic approxima- tion reveals the role of interaction overlap on the spread of complex contagions on higher-order networks.Phys. Rev. Lett.132, 077401 (2024)

    2024

  5. [13]

    Ferraz de Arruda, A

    G. Ferraz de Arruda, A. Aleta, and Y. Moreno, Conta- gion dynamics on higher-order networks.Nat. Rev. Phys. 6(8), 468–482 (2024)

    2024

  6. [14]

    Di Gaetano, G

    L. Di Gaetano, G. Carugno, F. Battiston, and F. Coghi, Dynamical fluctuations of random walks in higher-order networks.Phys. Rev. Lett.133, 107401 (2024). 10

    2024

  7. [15]

    P. S. Skardal and A. Arenas, Abrupt desynchronization and extensive multistability in globally coupled oscillator simplexes.Phys. Rev. Lett.122, 248301 (2019)

    2019

  8. [16]

    A. P. Mill´ an, J. J. Torres, and G. Bianconi, Explo- sive higher-order kuramoto dynamics on simplicial com- plexes.Phys. Rev. Lett.124, 218301 (2020)

    2020

  9. [17]

    Zhang, M

    Y. Zhang, M. Lucas, and F. Battiston, Higher-order in- teractions shape collective dynamics differently in hyper- graphs and simplicial complexes.Nat. Commun.14(1), 1605 (2023)

    2023

  10. [18]

    Alvarez-Rodriguez, F

    U. Alvarez-Rodriguez, F. Battiston, G. F. de Arruda, Y. Moreno, M. Perc, and V. Latora, Evolutionary dy- namics of higher-order interactions in social networks. Nat. Hum. Behav.5, 586–585 (2021)

    2021

  11. [19]

    Civilini, O

    A. Civilini, O. Sadekar, F. Battiston, J. G´ omez- Garde˜ nes, and V. Latora, Explosive cooperation in so- cial dilemmas on higher-order networks.Phys. Rev. Lett. 132(16), 167401 (2024)

    2024

  12. [20]

    D. A. Spielman and N. Srivastava, Graph sparsification by effective resistances. InProceedings of the Fortieth An- nual ACM Symposium on Theory of Computing563–568 (2008)

    2008

  13. [21]

    D. A. Spielman and S.-H. Teng, Spectral sparsification of graphs.SIAM Journal on Computing40, 981–1025 (2011)

    2011

  14. [22]

    Tumminello, S

    M. Tumminello, S. Micciche, F. Lillo, J. Piilo, and R. N. Mantegna, Statistically validated networks in bipartite complex systems.PloS One6, e17994 (2011)

    2011

  15. [23]

    Dianati, Unwinding the hairball graph: Pruning algo- rithms for weighted complex networks.Phys

    N. Dianati, Unwinding the hairball graph: Pruning algo- rithms for weighted complex networks.Phys. Rev. E93, 012304 (2016)

    2016

  16. [24]

    Casiraghi, V

    G. Casiraghi, V. Nanumyan, I. Scholtes, and F. Schweitzer, From relational data to graphs: In- ferring significant links using generalized hypergeometric ensembles. InInternational conference on social infor- matics111–120 (2017)

    2017

  17. [25]

    M. ´A. Serrano, M. Bogun´ a, and A. Vespignani, Extract- ing the multiscale backbone of complex weighted net- works.Proc. Natl. Acad. Sci. U.S. A106, 6483–6488 (2009)

    2009

  18. [26]

    Marcaccioli and G

    R. Marcaccioli and G. Livan, A P´ olya urn approach to information filtering in complex networks.Nat. Commun. 10, 745 (2019)

    2019

  19. [27]

    Kirkley, Fast nonparametric inference of network backbones for weighted graph sparsification.Phys

    A. Kirkley, Fast nonparametric inference of network backbones for weighted graph sparsification.Phys. Rev. X15, 031013 (2025)

    2025

  20. [28]

    Grady, C

    D. Grady, C. Thiemann, and D. Brockmann, Robust clas- sification of salient links in complex networks.Nat. Com- mun.3, 864 (2012)

    2012

  21. [29]

    Rajeh, M

    S. Rajeh, M. Savonnet, E. Leclercq, and H. Cherifi, Modularity-based backbone extraction in weighted com- plex networks. InInternational Conference on Network Science67–79 (2022)

    2022

  22. [30]

    Z. P. Neal, An R package to extract network backbones. PloS one17, e0269137 (2022)

    2022

  23. [31]

    Yassin, A

    A. Yassin, A. Haidar, H. Cherifi, H. Seba, and O. Togni, An evaluation tool for backbone extraction techniques in weighted complex networks.Sci. Rep.13, 17000 (2023)

    2023

  24. [32]

    Musciotto, F

    F. Musciotto, F. Battiston, and R. N. Mantegna, Detect- ing informative higher-order interactions in statistically validated hypergraphs.Commun. Phys.4, 218 (2021)

    2021

  25. [33]

    Musciotto, F

    F. Musciotto, F. Battiston, and R. Mantegna, Identifying maximal sets of significantly interacting nodes in higher- order networks. arXiv:2209.12712 (2022)

    2022 arXiv

  26. [34]

    N. W. Landry, I. Amburg, M. Shi, and S. G. Aksoy, Filtering higher-order datasets.J. Phys. Complex.5(1), 015006 (2024)

    2024

  27. [35]

    Ceria and F

    A. Ceria and F. W. Takes, The relevance of higher-order ties.EPJ Data Science14(1), 62 (2025)

    2025

  28. [36]

    Kirkley, H

    A. Kirkley, H. Felippe, and F. Battiston, Structural re- ducibility of hypergraphs.Phys. Rev. Lett.135, 247401 (2025)

    2025

  29. [37]

    Zhang, A

    S. Zhang, A. Ceria, and H. Wang, Diffusion backbone of temporal higher-order networks. arXiv:2412.12856 (2024)

    2024 arXiv

  30. [38]

    Lucas, L

    M. Lucas, L. Gallo, A. Ghavasieh, F. Battiston, and M. De Domenico, Reducibility of higher-order networks from dynamics.Nat. Commun.17, 1551 (2026)

    2026

  31. [39]

    Rissanen, Modeling by the shortest data description

    J. Rissanen, Modeling by the shortest data description. Automatica14, 465–471 (1978)

    1978

  32. [40]

    T. P. Peixoto, Nonparametric Bayesian inference of the microcanonical stochastic block model.Phys. Rev. E95, 012317 (2017)

    2017

  33. [41]

    T. P. Peixoto, Bayesian stochastic blockmodeling. In P. Doreian, V. Batagelj, and A. Ferligoj (eds.),Advances in Network Clustering and Blockmodeling, 289–332, Wi- ley, New York (2019)

    2019

  34. [42]

    Kirkley, Spatial regionalization based on optimal information compression.Commun

    A. Kirkley, Spatial regionalization based on optimal information compression.Commun. Phys.5(1), 1–10 (2022)

    2022

  35. [43]

    Morel-Balbi and A

    S. Morel-Balbi and A. Kirkley, Bayesian regionalization of urban mobility networks.Phys. Rev. Res.6(3), 033307 (2024)

    2024

  36. [44]

    H´ ebert-Dufresne, J.-G

    L. H´ ebert-Dufresne, J.-G. Young, A. Daniels, A. Kirkley, and A. Allard, Network compression with configuration models and the minimum description length.Phys. Rev. E110, 034305 (2024)

    2024

  37. [45]

    T. P. Peixoto and A. Kirkley, Implicit models, latent compression, intrinsic biases, and cheap lunches in com- munity detection.Phys. Rev. E.108(2), 024309 (2023)

    2023

  38. [46]

    Coupette and J

    C. Coupette and J. Vreeken, Graph similarity descrip- tion: How are these graphs similar? InProceedings of the 27th ACM SIGKDD Conference on Knowledge Dis- covery & Data Mining, 185–195 (2021)

    2021

  39. [47]

    Coupette, S

    C. Coupette, S. Dalleiger, and J. Vreeken, Differ- entially describing groups of graphs. InProceedings of the AAAI Conference on Artificial Intelligence36, 3959–3967 (2022)

    2022

  40. [48]

    Kirkley, A

    A. Kirkley, A. Rojas, M. Rosvall, and J.-G. Young, Com- pressing network populations with modal networks reveal structural diversity.Commun. Phys.6, 148 (2023)

    2023

  41. [49]

    Felippe, F

    H. Felippe, F. Battiston, and A. Kirkley, Network mu- tual information measures for graph similarity.Commun. Phys.7(1), 335 (2024)

    2024

  42. [50]

    Jerdee, A

    M. Jerdee, A. Kirkley, and M. Newman, Mutual informa- tion and the encoding of contingency tables.Phys. Rev. E.110(6), 064306 (2024)

    2024

  43. [51]

    Q. F. Lotito, F. Musciotto, A. Montresor, and F. Battis- ton, Higher-order motif analysis in hypergraphs.Com- mun. Phys.5(1), 79 (2022)

    2022

  44. [52]

    LaRock and R

    T. LaRock and R. Lambiotte, Encapsulation structure and dynamics in hypergraphs.J. Phys. Complex.4(4), 045007 (2023)

    2023

  45. [53]

    N. W. Landry, J.-G. Young, and N. Eikmeier, The sim- pliciality of higher-order networks.EPJ Data Sci.13(1), 17 (2024). 11

    2024

  46. [54]

    Gallo, L

    L. Gallo, L. Lacasa, V. Latora, and F. Battiston, Higher- order correlations reveal complex memory in temporal hypergraphs.Nat. Commun.15(1), 4754 (2024)

    2024

  47. [55]

    D. J. MacKay,Information theory, inference and learning algorithms. Cambridge University Press (2003)

    2003

  48. [56]

    Felippe, A

    H. Felippe, A. Kirkley, and F. Battiston, Information the- ory for hypergraph similarity. arXiv:2510.27411 (2025)

    2025 arXiv

  49. [57]

    M. E. Newman, G. T. Cantwell, and J.-G. Young, Im- proved mutual information measure for clustering, classi- fication, and community detection.Phys. Rev. E.101(4), 042304 (2020)

    2020

  50. [58]

    M. H. Hansen and B. Yu, Model selection and the princi- ple of minimum description length.Journal of the Amer- ican Statistical Association96, 746–774 (2001)

    2001

  51. [59]

    Clauset, M

    A. Clauset, M. E. Newman, and C. Moore, Finding com- munity structure in very large networks.Phys. Rev. E 70, 066111 (2004)

    2004

  52. [60]

    V. D. Blondel, J.-L. Guillaume, R. Lambiotte, and E. Lefebvre, Fast unfolding of communities in large net- works.Journal of Statistical Mechanics: Theory and Ex- perimentP10008 (2008)

    2008

  53. [61]

    Kirkley, Identifying hubs in directed networks.Phys

    A. Kirkley, Identifying hubs in directed networks.Phys. Rev. E.109, 034310 (2024)

    2024

  54. [62]

    Q. F. Lotito, L. Betti, B. Nortier, A. Montresor, and F. Battiston, Hypergraphx-data: a repository for higher- order network data.J. Complex. Netw.(2026)

    2026

  55. [63]

    Young, G

    J.-G. Young, G. Petri, and T. P. Peixoto, Hypergraph reconstruction from network data.Commun. Phys.4, 135 (2021)

    2021

  56. [64]

    Lizotte, J.-G

    S. Lizotte, J.-G. Young, and A. Allard, Hypergraph re- construction from uncertain pairwise observations.Sci. Rep.13, 21364 (2023)

    2023

  57. [65]

    A. E. Wegner, Subgraph covers: an information-theoretic approach to motif analysis in networks.Phys. Rev. X4, 041026 (2014)

    2014

  58. [66]

    A. E. Wegner and S. C. Olhede, Nonparametric inference of higher order interaction patterns in networks.Com- mun. Phys.7, 258 (2024)

    2024

  59. [67]

    Kirkley, Inference of dynamic hypergraph representa- tions in temporal interaction data.Phys

    A. Kirkley, Inference of dynamic hypergraph representa- tions in temporal interaction data.Phys. Rev. E.109(5), 054306 (2024)

    2024

  60. [68]

    Q. F. Lotito, M. Contisciani, C. De Bacco, L. Di Gaetano, L. Gallo, A. Montresor, F. Musciotto, N. Ruggeri, and F. Battiston, Hypergraphx: a library for higher-order network analysis.J. Complex. Netw.11(3), cnad019 (2023)

    2023

  61. [69]

    T. M. Cover and J. A. Thomas,Elements of Information Theory. John Wiley & Sons (2012)

    2012

  62. [70]

    node” and connecting all pairs of adjacent “nodes

    E. Vasilyeva, M. Romance, I. Samoylenko, K. Kovalenko, D. Musatov, A. M. Raigorodskii, and S. Boccaletti Distances in weighted higher-order networks.Commun. Phys.9(178) (2026). 12 Appendix A: Proof of RMI properties The RMI in Eq. (11) can be equivalently written in terms of t...

    2026

Pith tools

Reviewed July 12, 2026 · model on record in the stance chip above.