Pith. sign in

REVIEW 2 major objections 5 minor 1 cited by

Immunization on Temporal Higher-Order Networks

T0 review · 2 major / 5 minor · reviewed 2026-07-14 · grok-4.5

Pith's one-line read On temporal higher-order networks, immunization success depends on how large the outbreak already is, and a prevalence-weighted ranking of nodes clears infection more efficiently than standard activity heuristics.

desk verdict Real contribution on immunization under temporal higher-order contagion; HIC ranking is useful but rests on a homogeneous-infection closure that is weakest exactly where higher-order effects dominate. read the letter →

arxiv 2607.10171 v1 pith:X24V4RU2 submitted 2026-07-11 physics.soc-ph

classification physics.soc-ph
keywords higher-orderimmunizationinfectioncontributioncontagiondiscontinuousphasetransitionsegocentricsamplingstrategiestemporalhypergraphsactivity-drivenmodelbistability
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

This paper studies how to stop contagion when contacts are both group-based and constantly changing. On such temporal higher-order networks the fraction of people you need to immunize is not a single fixed number: as that fraction rises, prevalence can jump discontinuously to extinction, and for a range of fractions both extinction and a large endemic state are possible, so the outcome hinges on the prevalence at the moment of intervention. That dependence is absent from ordinary pairwise temporal networks. From the same mean-field description the authors derive a High Infection Contribution score that weights each person's pairwise and group activity by the current prevalence, and show that ranking people by this score yields the lowest eradication threshold among the global strategies they test. When only local samples of contacts are available, pairwise-based sampling works better early in an outbreak while group-based sampling works better later; a simple two-stage rule that switches between them therefore improves on random immunization. The same qualitative picture appears on a real face-to-face contact dataset.

What carries the argument

The High Infection Contribution (HIC) score IC(a) = 2β₁a⁽¹⁾ + 3β₂ρ₀(1−ω)a⁽²⁾, obtained by minimizing the one-step post-immunization rise in prevalence under a homogeneous-infection closure of the higher-order activity-driven mean-field equations; it supplies both the global ranking rule and the diagnostic that explains when pairwise versus higher-order targeting is superior.

What would settle it

On the same higher-order activity-driven networks, replace the homogeneous pre-immunization infection profile with a strongly activity-biased one (or measure the actual profile from full simulations) and check whether HIC still yields the lowest eradication threshold and whether the HA–PA and EPS–EHS crossovers still occur at the predicted prevalence.

Watch

Extended reading notes

Core claim

Immunization on temporal higher-order networks produces bistability and discontinuous (hybrid) transitions in steady-state prevalence as the immunized fraction varies, so the immunization threshold itself depends on pre-immunization prevalence. Ranking nodes by the infection-contribution score IC(a) = 2β₁a⁽¹⁾ + 3β₂ρ₀(1−ω)a⁽²⁾ therefore outperforms total-, pairwise- and higher-order-activity heuristics, while among local (egocentric) strategies the better choice itself switches from pairwise to higher-order sampling as prevalence rises.

Load-bearing premise

The derivation treats infection as evenly spread across activity classes just before immunization; if high-activity people are already far more infected than average, the ranking and the claimed crossovers can change.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

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 studies immunization of a nonlinear higher-order contagion process on temporal hypergraphs generated by the higher-order activity-driven (HOAD) model. Using a mean-field description (Eqs. 1–3), it shows that as the immunized fraction ω varies, the steady-state prevalence exhibits discontinuous (hybrid) transitions and bistability, so the eradication threshold ω_c depends on the pre-immunization prevalence ρ₀—unlike temporal pairwise networks. Motivated by that dependence, the authors derive a High Infection Contribution (HIC) ranking IC(a)=2β₁a⁽¹⁾+3β₂ρ₀(1−ω)a⁽²⁾ that minimizes one-step post-immunization growth under a homogeneous-profile closure, and show that HIC yields the lowest ω_c among the global heuristics tested (TA, HA, PA, R). They further introduce egocentric sampling strategies (EPS, EHS, EBS) based on local pairwise/triadic counts and a two-stage rule (TES) that switches with ρ₀, and reproduce the qualitative phenomenology on an augmented SocioPatterns contact sequence.

Significance. If the results hold, the work supplies a concrete, prevalence-aware immunization theory for temporal higher-order systems and a practical ranking (HIC) that systematically outperforms standard activity heuristics. The early-stage threshold (Eq. 4), the fixed-point stability procedure for finite ρ₀, and the closed-form egocentric nomination intensities (Eqs. 11–12, 49) are carefully derived and track Monte Carlo simulations; the SocioPatterns validation shows the same discontinuous/bistable structure and strategy crossovers. These elements are useful for epidemic and misinformation control when group interactions are both higher-order and time-varying, and they cleanly separate the dynamical novelty (ρ₀-dependent thresholds) from the design of deployable local strategies.

major comments (2)
  1. Methods §4.4–4.5 (Eqs. 5, 25–26, 28–30, 32–37): HIC and the equal-threshold IC comparison that explain the HA–PA and EPS–EHS crossovers are derived by closing one-step growth under the homogeneous pre-immunization profile i^{t⁻₀}_a ≈ ρ₀ n_a. Under the same nonlinear higher-order infection rule that produces bistability, high-activity classes are preferentially infected before t₀, so the true profile is concentrated on large a—most strongly at high ρ₀, precisely where HA/EHS are claimed to overtake PA/EPS. The paper should either (i) recompute IC and the crossover loci with the actual pre-immunization fixed-point profile i^{t⁻₀}_a obtained from Eq. (20), or (ii) quantify the ranking error of the homogeneous closure against that profile across the ρ₀ range of Figs. 2–3. Without this check the absolute optimality of HIC and the IC-based explanation of the crossovers remain incompletely supp
  2. §2.3 and Methods §4.3: Instantaneous thinning of the infected density at t₀ (Eq. 19) is used both for the theoretical thresholds and for the HIC derivation. Real immunization (vaccination, isolation) acts with a delay and does not instantly remove already-infected individuals from the infectious pool. The paper should report at least one delayed-immunization or gradual-rollout protocol (e.g., continuous removal of a fraction of S and I over a finite window after t₀) and show whether the discontinuous transitions, the ρ₀-dependence of ω_c, and the HIC ranking order survive. If they do not, the practical claim that HIC is the preferred strategy needs to be qualified.
minor comments (5)
  1. Fig. 1b–c: the hybrid critical scalings |ρ*−ρ*_ωL|∝|ω−ω_L| and |ρ*−ρ*_ωU|∝|ω−ω_U|^{0.5} are stated without error bars or fit ranges; a short table of fitted exponents and residual norms would make the hybrid claim more transparent.
  2. §2.4 / Methods §4.6: the probe fraction ϕ and window ΔT are free parameters of the egocentric strategies but are not systematically varied in the main figures; a brief sensitivity panel (or SI note) would clarify robustness of the EPS–EHS crossover.
  3. Methods §4.7: the empirical network is expanded by a factor of 100 to N=15500. The text should state whether the activity-rate ranks (and therefore HIC order) are preserved under this augmentation, or report the same strategy comparison on the original N.
  4. Notation: ρ₀, ρ^{+}_{0}, ρ̃_M and ω_L / ω_U / ω_c appear with slightly different subscripts across the abstract, §2.1 and Methods; a single consistent glossary would help.
  5. Discussion: the independence of a⁽¹⁾ and a⁽²⁾ is listed as a limitation; a one-sentence remark on how a modest positive correlation would shift IC weights would be useful for readers applying the method to empirical data.

Circularity Check

0 steps flagged · score 1.0 of 10

No load-bearing circularity: thresholds and HIC ranking follow from mean-field fixed-point analysis plus a one-step growth approximation, then are checked against independent Monte Carlo and empirical runs.

full rationale

The core phenomenology (discontinuous/hybrid transitions and bistability of ρ* vs ω, with ωc depending on pre-immunization ρ0) is obtained by solving the closed mean-field map (Eqs. 1–3) for fixed points and linear stability (Methods 4.2–4.3), then confirmed by direct simulation of the HOAD process; nothing is fitted to the target curves. HIC is constructed by minimizing the instantaneous post-immunization growth Δρ+0 under the homogeneous-profile closure i^{t−0}_a ≈ ρ0 na (Methods 4.4, Eqs. 25–30); the resulting IC score is therefore a model-based ranking heuristic, not a tautological re-labeling of the observed ωc. The equal-threshold condition ⟨IC⟩imm^X = ⟨IC⟩imm^Y (Methods 4.5) is likewise an algebraic consequence of the same closure and is used only to interpret the HA–PA / EPS–EHS crossovers already measured from the full dynamics. Egocentric nomination intensities F_X(a) are derived from expected co-occurrence counts under the HOAD generative process (Methods 4.6) and then inserted into the same mean-field equations. Self-citations are to the external HOAD construction and standard simplicial-contagion literature; none supply a uniqueness theorem or ansatz that forces the present claims. The homogeneous-profile approximation may be inaccurate at high ρ0, but that is a modeling assumption, not a circular reduction of prediction to input. Hence the derivation chain is self-contained against the paper’s own simulations and the SocioPatterns validation.

Assumptions & free parameters 6 free parameters · 6 assumptions · 2 invented entities

The central claims live inside a standard mean-field higher-order contagion setup on HOAD temporal hypergraphs. Load-bearing modeling choices (nonlinear all-infected group infection, activity independence, instantaneous immunization, homogeneous pre-immunization infection for HIC) and free rate/distribution parameters set the quantitative thresholds; HIC/IC and the named egocentric strategies are the main invented constructs. No new physical entity is postulated beyond these modeling objects.

free parameters (6)
  • β₁ (pairwise infection rate)
    Chosen for simulations (e.g. 0.0034 synthetic; 0.008 empirical); sets relative weight of pairwise terms in thresholds and IC.
  • β₂ (higher-order infection rate)
    Chosen (e.g. 0.06 synthetic; 0.05 empirical); controls strength of discontinuous/bistable regime and HA/EHS advantage at high ρ₀.
  • μ (recovery rate)
    Fixed at 0.001 in reported runs; scales the epidemic time unit and enters the early threshold condition.
  • Activity distribution exponents and means ⟨a⁽¹⟩⟩, ⟨a⁽²⟩⟩
    Power-law −2.1 with ⟨a⁽¹⟩⟩=0.13, ⟨a⁽²⟩⟩=0.03 in synthetic experiments; shape heterogeneity that targeting exploits.
  • Probe fraction ϕ and observation window ΔT
    Control how much local structure egocentric strategies see; enter q_a ≈ 1−exp[−ϕ F_X(a)].
  • Empirical network augmentation factor (×100 → N=15500)
    Hand choice to reduce finite-size noise while preserving temporal pattern; can affect absolute thresholds.
assumptions (6)
  • domain assumption Mean-field activity-class dynamics close the contagion on HOAD hypergraphs (Eqs. 1–3).
    Standard in activity-driven epidemic theory; neglects dynamical correlations and finite-N fluctuations.
  • domain assumption Higher-order infection: a susceptible in an (m+1)-hyperedge is infected at rate β_m only if all other members are infected.
    Nonlinear simplicial/hypergraph mechanism from prior contagion literature; source of bistability.
  • domain assumption Pairwise and higher-order activity components are drawn independently; joint density factorizes.
    Stated in Methods §4.1; Discussion notes correlations may change prioritization.
  • ad hoc to paper Immunization instantly sets r_a = q_a n_a and thins infected density to (1−q_a)i_a at t₀.
    Explicit tractability assumption in §2.1; real antibody lag and gradual rollout differ.
  • ad hoc to paper Pre-immunization infection is approximately homogeneous across activity classes when deriving HIC (i_a ≈ ρ₀ n_a).
    Methods §4.4; closes one-step growth to the IC score used for ranking.
  • standard math Linear stability and fixed-point classification determine ω_c(ρ₀) in the mean-field map.
    Standard dynamical-systems threshold analysis (Methods §4.2–4.3).
invented entities (2)
  • Infection Contribution (IC) score / High Infection Contribution (HIC) strategy
    purpose: Rank nodes by estimated contribution to one-step post-immunization prevalence growth and immunize highest-IC nodes.
    Defined from minimizing J(q) under the homogeneous closure; not an independently measured biological quantity.
  • Egocentric strategies EHS, EPS, EBS and two-stage TES
    purpose: Nominate immunization targets from local pairwise/triadic counts observed by probes, with TES switching by prevalence.
    Operational sampling rules introduced here; performance is model- and data-dependent.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Immunization on Temporal Higher-Order Networks." pith.science (2026). https://pith.science/paper/X24V4RU2

@misc{pith2026260710171,
  author       = {Pith},
  title        = {Pith review of: Immunization on Temporal Higher-Order Networks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/X24V4RU2}},
  note         = {Machine review of arXiv:2607.10171}
}
read the original abstract

Network immunization is a powerful tool for controlling contagion processes ranging from infectious diseases to misinformation diffusion. While prior works have focused on pairwise or static networks, immunization dynamics in temporal higher-order networks remain poorly understood. Here, we introduce immunization strategies and develop a theoretical framework tailored for such temporal systems. Firstly, we reveal bistability and discontinuous transitions in prevalence as the immunization fraction varies. This implies that immunization effectiveness depends on the initial prevalence, marking a fundamental departure from pairwise networks. Building on this prevalence-dependent behavior, we propose the High Infection Contribution (HIC) strategy, demonstrating its superior performance over all evaluated heuristic strategies. Furthermore, we introduce egocentric strategies by leveraging solely local observations. Notably, the optimal egocentric strategy shifts with the contagion prevalence. Our work advances the understanding of network immunization, paving the way for effective contagion control in temporal higher-order networks.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

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

  1. Adaptive higher-order contagion of harmful information with platform-induced group dissolution and individual rewiring

    physics.soc-ph 2026-08 conditional novelty 6.0 of 10

    In an adaptive higher-order contagion model, group dissolution eradicates harmful information only below a critical infection rate; above it, dissolution backfires and increases prevalence.

Reference graph

Works this paper leans on

47 extracted references · 1 linked inside Pith · cited by 1 Pith paper

  1. [1]

    & Jansen, V

    Funk, S., Salath´ e, M. & Jansen, V. A. Modelling the influence of human behaviour on the spread of infectious diseases: a review.Journal of the Royal Society Interface7, 1247–1256 (2010)

  2. [2]

    Del Vicario, M.et al.The spreading of misinformation online.Proceedings of the national academy of Sciences113, 554–559 (2016)

  3. [3]

    M.et al.The science of fake news.Science359, 1094–1096 (2018)

    Lazer, D. M.et al.The science of fake news.Science359, 1094–1096 (2018)

  4. [4]

    Frontiers in public health9, 675687 (2021)

    Fu, P.et al.Propagation model of panic buying under the sudden epidemic. Frontiers in public health9, 675687 (2021)

  5. [5]

    & Gentzkow, M

    Allcott, H. & Gentzkow, M. Social media and fake news in the 2016 election. Journal of economic perspectives31, 211–236 (2017)

  6. [6]

    & Fernandes, M

    Azzimonti, M. & Fernandes, M. Social media networks, fake news, and polarization.European journal of political economy76, 102256 (2023)

  7. [7]

    & De Domenico, M

    Gallotti, R., Valle, F., Castaldo, N., Sacco, P. & De Domenico, M. Assessing the risks of ‘infodemics’ in response to covid-19 epidemics.Nature human behaviour 4, 1285–1293 (2020)

  8. [8]

    Kouzy, R.et al.Coronavirus goes viral: quantifying the covid-19 misinformation epidemic on twitter.Cureus12(2020)

Show all 47 references
  1. [9]

    J., Yoon, C

    Holme, P., Kim, B. J., Yoon, C. N. & Han, S. K. Attack vulnerability of complex networks.Physical review E65, 056109 (2002)

  2. [10]

    & Vespignani, A

    Pastor-Satorras, R. & Vespignani, A. Immunization of complex networks.Physical review E65, 036104 (2002)

  3. [11]

    & Ben-Avraham, D

    Cohen, R., Havlin, S. & Ben-Avraham, D. Efficient immunization strategies for computer networks and populations.Physical review letters91, 247901 (2003)

  4. [12]

    Nature physics6, 888–893 (2010)

    Kitsak, M.et al.Identification of influential spreaders in complex networks. Nature physics6, 888–893 (2010)

  5. [13]

    & Vespignani, A

    Liu, S., Perra, N., Karsai, M. & Vespignani, A. Controlling contagion processes in activity driven networks.Physical review letters112, 118702 (2014)

  6. [14]

    Physical Review E—Statistical, Nonlinear, and Soft Matter Physics84, 016101 (2011)

    Van Mieghem, P.et al.Decreasing the spectral radius of a graph by link removals. Physical Review E—Statistical, Nonlinear, and Soft Matter Physics84, 016101 (2011)

  7. [15]

    M., Mihaljev, T., Havlin, S

    Schneider, C. M., Mihaljev, T., Havlin, S. & Herrmann, H. J. Suppress- ing epidemics with a limited amount of immunization units.Physical Review 21 E—Statistical, Nonlinear, and Soft Matter Physics84, 061911 (2011)

  8. [16]

    Bishop, A. N. & Shames, I. Link operations for slowing the spread of disease in complex networks.EPL (Europhysics Letters)95, 18005 (2011)

  9. [17]

    T., Arenas, A

    Matamalas, J. T., Arenas, A. & G´ omez, S. Effective approach to epidemic con- tainment using link equations in complex networks.Science advances4, eaau4212 (2018)

  10. [18]

    R., Gleich, D

    Benson, A. R., Gleich, D. F. & Leskovec, J. Higher-order organization of complex networks.Science353, 163–166 (2016)

  11. [19]

    Grilli, J., Barab´ as, G., Michalska-Smith, M. J. & Allesina, S. Higher-order inter- actions stabilize dynamics in competitive network models.Nature548, 210–213 (2017)

  12. [20]

    Mayfield, M. M. & Stouffer, D. B. Higher-order interactions capture unexplained complexity in diverse communities.Nature ecology & evolution1, 0062 (2017)

  13. [21]

    Nature Physics17, 1093–1098 (2021)

    Battiston, F.et al.The physics of higher-order interactions in complex systems. Nature Physics17, 1093–1098 (2021)

  14. [22]

    Boccaletti, S.et al.The structure and dynamics of networks with higher order interactions.Physics Reports1018, 1–64 (2023)

  15. [23]

    & Latora, V

    Iacopini, I., Petri, G., Barrat, A. & Latora, V. Simplicial models of social contagion.Nature communications10, 2485 (2019)

  16. [24]

    F., Petri, G

    de Arruda, G. F., Petri, G. & Moreno, Y. Social contagion models on hypergraphs. Physical Review Research2, 023032 (2020)

  17. [25]

    & Bianconi, G

    St-Onge, G., Sun, H., Allard, A., H´ ebert-Dufresne, L. & Bianconi, G. Universal nonlinear infection kernel from heterogeneous exposure on higher-order networks. Physical review letters127, 158301 (2021)

  18. [26]

    St-Onge, G.et al.Influential groups for seeding and sustaining nonlinear contagion in heterogeneous hypergraphs.Communications Physics5, 25 (2022)

  19. [27]

    Ferraz de Arruda, G., Petri, G., Rodriguez, P. M. & Moreno, Y. Multistability, intermittency, and hybrid transitions in social contagion models on hypergraphs. Nature communications14, 1375 (2023)

  20. [28]

    & Moreno, Y

    Ferraz de Arruda, G., Aleta, A. & Moreno, Y. Contagion dynamics on higher- order networks.Nature Reviews Physics6, 468–482 (2024)

  21. [29]

    Landry, N. W. & Restrepo, J. G. The effect of heterogeneity on hypergraph contagion models.Chaos: An Interdisciplinary Journal of Nonlinear Science30 (2020). 22

  22. [30]

    & Goh, K.-I

    Kim, J., Lee, D.-S. & Goh, K.-I. Contagion dynamics on hypergraphs with nested hyperedges.Physical Review E108, 034313 (2023)

  23. [31]

    & Goh, K.-I

    Kim, J.-H. & Goh, K.-I. Higher-order components dictate higher-order contagion dynamics in hypergraphs.Physical review letters132, 087401 (2024)

  24. [32]

    & Arenas, A

    Burgio, G., G´ omez, S. & Arenas, A. Triadic approximation reveals the role of interaction overlap on the spread of complex contagions on higher-order networks. Physical Review Letters132, 077401 (2024)

  25. [33]

    & G´ omez-Garde˜ nes, J

    Malizia, F., Lamata-Ot´ ın, S., Frasca, M., Latora, V. & G´ omez-Garde˜ nes, J. Hyperedge overlap drives explosive transitions in systems with higher-order interactions.Nature communications16, 555 (2025)

  26. [34]

    & Battiston, F

    Chowdhary, S., Kumar, A., Cencetti, G., Iacopini, I. & Battiston, F. Simplicial contagion in temporal higher-order networks.Journal of Physics: Complexity2, 035019 (2021)

  27. [35]

    Han, Z.et al.Probabilistic activity driven model of temporal simplicial net- works and its application on higher-order dynamics.Chaos: An Interdisciplinary Journal of Nonlinear Science34(2024)

  28. [36]

    & Barrat, A

    Iacopini, I., Karsai, M. & Barrat, A. The temporal dynamics of group interactions in higher-order social networks.Nature Communications15, 7391 (2024)

  29. [37]

    & Battiston, F

    Gallo, L., Lacasa, L., Latora, V. & Battiston, F. Higher-order correlations reveal complex memory in temporal hypergraphs.Nature Communications15, 4754 (2024)

  30. [38]

    & Tang, S

    Liu, L., Zheng, H., Han, Z., Wang, X. & Tang, S. Higher-order network adaptivity: co-evolution of higher-order structure and spreading dynamics.arXiv preprint arXiv:2508.15445(2025)

  31. [39]

    & H´ ebert-Dufresne, L

    Burgio, G., St-Onge, G. & H´ ebert-Dufresne, L. Characteristic scales and adaptation in higher-order contagions.Nature Communications16, 4589 (2025)

  32. [40]

    Effective epidemic containment strategy in hypergraphs.Physical Review Research3, 033282 (2021)

    Jhun, B. Effective epidemic containment strategy in hypergraphs.Physical Review Research3, 033282 (2021)

  33. [41]

    Li, W.et al.Immunization strategies for simplicial irreversible epidemic on simplicial complex.Frontiers in Physics10, 1018844 (2022)

  34. [42]

    & Wang, W

    Nie, Y., Su, S., Lin, T., Liu, Y. & Wang, W. Voluntary vaccination on hypergraph. Communications in Nonlinear Science and Numerical Simulation127, 107594 (2023)

  35. [43]

    & Tan, Y

    Yang, J.-X., Xiao, Y.-L., Zhang, J. & Tan, Y. Centrality measures to control epidemic spreading on uniform hypergraphs or hypernetworks.Communications 23 in Nonlinear Science and Numerical Simulation109368 (2025)

  36. [44]

    & Starnini, M

    Di Gaetano, L., Battiston, F. & Starnini, M. Percolation and topological prop- erties of temporal higher-order networks.Physical Review Letters132, 037401 (2024)

  37. [45]

    Journal of theoretical biology271, 166–180 (2011)

    Isella, L.et al.What’s in a crowd? analysis of face-to-face behavioral networks. Journal of theoretical biology271, 166–180 (2011)

  38. [46]

    & Perra, N

    Pozzana, I., Sun, K. & Perra, N. Epidemic spreading on activity-driven networks with attractiveness.Physical Review E96, 042310 (2017)

  39. [47]

    & Patania, A

    Young, J.-G., Petri, G., Vaccarino, F. & Patania, A. Construction of and efficient sampling from the simplicial configuration model.Physical Review E96, 032312 (2017). 24

Pith tools

Reviewed July 14, 2026 · model on record in the stance chip above.