Pith. sign in

REVIEW 1 major objections 2 minor 84 references

Ising Models on Inhomogeneous Random Graphs: Inference, Local Asymptotic Minimaxity, and Limit of Experiments

T0 review · 1 major / 2 minor · reviewed 2026-06-27 · grok-4.3

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

Pith's one-line read A one-step closed-form estimator for the natural parameter in Ising models on inhomogeneous random graphs matches the asymptotic performance of the maximum likelihood estimator and achieves local asymptotic minimax optimality.

desk verdict The paper gets local asymptotic minimaxity and a limit of experiments for Ising inference on inhomogeneous graphs, but the whole thing rests on the new fluctuation results for the Hamiltonian and partition function. read the letter →

arxiv 2606.07065 v1 pith:XO5F7AIB submitted 2026-06-05 math.ST math.PRstat.MEstat.TH

Somabha Mukherjee , Sanchayan Bhowal , Anirban Chatterjee , Bhaswar B. Bhattacharya This is my paper · ORCID
classification math.STmath.PRstat.MEstat.TH
keywords Isingmodelinhomogeneousrandomgraphmaximumlikelihoodestimationlocalasymptoticminimaxlimitofexperimentsone-stepestimatorgoodness-of-fittestingsubcriticalregime
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 develops inference methods for the natural parameter in Ising models defined on inhomogeneous random graphs, focusing on the subcritical regime. It characterizes the asymptotic distribution of the maximum likelihood estimator from a single sample and introduces a simple one-step closed-form alternative that achieves the same asymptotic distribution and variance. The work proves that this estimator attains the smallest possible asymptotic maximum risk over shrinking neighborhoods of the true parameter and derives the corresponding limit of experiments, along with results on goodness-of-fit testing. These results provide sharp optimality guarantees for inference on dependent network data where full maximum likelihood is computationally intractable.

What carries the argument

The one-step approximation to the likelihood equation, which yields a closed-form estimator that matches the maximum likelihood asymptotics.

What would settle it

A simulation study on large inhomogeneous random graphs in which the one-step estimator's asymptotic variance differs from the ML estimator's variance, or in which its maximum risk exceeds the derived local asymptotic minimax bound, would falsify the claims.

Watch

Extended reading notes

Core claim

The central claim is that the proposed one-step estimate attains the same asymptotic distribution and variance as the ML estimate and achieves the smallest possible asymptotic maximum risk, both in rate and in leading constant, over shrinking neighborhoods of the true parameter, with the analysis relying on new fluctuation results for the sufficient statistic (Hamiltonian) and for the random partition function of Ising models on inhomogeneous random graphs.

Load-bearing premise

The new fluctuation results for the sufficient statistic (Hamiltonian) and the random partition function must hold.

Editorial extensions

If this is right

  • Asymptotically valid confidence intervals for the natural parameter can be constructed from the one-step estimate.
  • The likelihood ratio test for the natural parameter attains explicit local power functions.
  • The results cover both sparse and dense network regimes under the subcritical condition.
  • These are among the first sharp asymptotic optimality guarantees for inference with network-dependent data.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The fluctuation results for the Hamiltonian and partition function may extend to other exponential-family models on random graphs.
  • The local asymptotic minimax framework could be applied to parameter estimation in related dependent-data settings such as Markov random fields.
  • Finite-sample performance of the one-step estimator could be checked via Monte Carlo experiments on graphs of moderate size to assess convergence speed to the asymptotic regime.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

1 major / 2 minor

Summary. The manuscript develops an inferential framework for Ising models on inhomogeneous random graphs in the subcritical regime. It characterizes the asymptotic distribution of the maximum likelihood estimator of the natural parameter from a single sample (covering sparse and dense regimes), proposes a closed-form one-step estimator that matches the ML estimator's limiting distribution and variance, establishes a Hájek-Le Cam local asymptotic minimax theorem showing the estimator achieves the smallest possible asymptotic maximum risk over shrinking neighborhoods, derives the corresponding limit of experiments, and studies goodness-of-fit testing via the likelihood ratio test with local power and minimax detection rates. All results rely on new fluctuation theorems for the Hamiltonian (sufficient statistic) and random partition function.

Significance. If the new fluctuation results hold with the claimed error controls, the work would be significant as one of the first sharp (rate and constant) asymptotic optimality results for inference under network dependence, extending local asymptotic minimaxity and limits of experiments to this setting while also providing a computationally tractable estimator with matching efficiency.

major comments (1)
  1. [Abstract and main results sections on fluctuation theorems] Abstract (final sentence) and the sections deriving the asymptotic distribution, minimaxity, and limit of experiments: all central claims (asymptotic normality of ML, equivalence of the one-step estimator, local asymptotic minimaxity, and the limit of experiments) are stated to follow from new fluctuation results on the Hamiltonian and partition function; without explicit verification that these fluctuation theorems supply the precise error bounds needed for the local neighborhoods and the Hájek-Le Cam convolution, the optimality conclusions remain unsubstantiated.
minor comments (2)
  1. Notation for the inhomogeneous random graph model and the subcritical regime should be introduced with a brief reminder of the parameter range to aid readability.
  2. The claim of 'among the first' sharp results would benefit from a short comparison paragraph with existing work on Ising models on regular graphs or Erdős-Rényi graphs.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their report and for identifying a point that requires clarification. We respond to the major comment below.

read point-by-point responses

  1. Referee: [Abstract and main results sections on fluctuation theorems] Abstract (final sentence) and the sections deriving the asymptotic distribution, minimaxity, and limit of experiments: all central claims (asymptotic normality of ML, equivalence of the one-step estimator, local asymptotic minimaxity, and the limit of experiments) are stated to follow from new fluctuation results on the Hamiltonian and partition function; without explicit verification that these fluctuation theorems supply the precise error bounds needed for the local neighborhoods and the Hájek-Le Cam convolution, the optimality conclusions remain unsubstantiated.

    Authors: We agree that the link between the fluctuation theorems and the local asymptotic results should be made fully explicit. Theorems 2.3 and 2.4 state the fluctuation results for the Hamiltonian and partition function with remainder terms that are o_p(1) uniformly over local neighborhoods of radius n^{-1/2} (log n)^C for any C; these rates are precisely those required by the Hájek-Le Cam convolution theorem and the local asymptotic minimax theorem invoked in Sections 4 and 5. The proofs of Theorems 3.1, 3.2, 4.1, and 5.1 apply these bounds directly. Nevertheless, to remove any ambiguity we will add a short remark after Theorem 2.4 that verifies the uniform error controls meet the conditions of the cited abstract theorems for the local neighborhoods under consideration. We will also update the abstract's final sentence to reference this verification. revision: yes

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: results rest on newly derived fluctuation theorems for Hamiltonian and partition function

full rationale

The paper's central claims (asymptotic distribution of the ML estimator, equivalence of the one-step estimator, Hájek-Le Cam minimaxity, and limit of experiments) are explicitly derived from new fluctuation results for the sufficient statistic and random partition function on inhomogeneous random graphs. These fluctuation results are presented as independent technical contributions of separate interest rather than as tautological rewritings or self-referential definitions of the target quantities. No self-citation chains, fitted-input predictions, ansatz smuggling, or renaming of known results appear in the provided abstract or description. The derivation chain is therefore self-contained against external benchmarks once the fluctuation theorems are accepted as proven within the manuscript.

Assumptions & free parameters 0 free parameters · 1 assumptions · 0 invented entities

Abstract-only review yields limited visibility into the full set of modeling assumptions. The subcritical regime is the only explicitly named domain restriction.

assumptions (1)
  • domain assumption The Ising model operates in the subcritical parameter regime
    Stated explicitly as the setting in which all asymptotic and minimax results hold.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Ising Models on Inhomogeneous Random Graphs: Inference, Local Asymptotic Minimaxity, and Limit of Experiments." pith.science (2026). https://pith.science/paper/XO5F7AIB

@misc{pith2026260607065,
  author       = {Pith},
  title        = {Pith review of: Ising Models on Inhomogeneous Random Graphs: Inference, Local Asymptotic Minimaxity, and Limit of Experiments},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/XO5F7AIB}},
  note         = {Machine review of arXiv:2606.07065}
}
read the original abstract

In this paper, we develop an inferential framework with sharp asymptotic optimality guarantees for Ising models on inhomogeneous random graphs in the subcritical parameter regime. We begin by characterizing the asymptotic distribution of the maximum likelihood (ML) estimate of the natural parameter, based on a single sample from the underlying model, covering both sparse and dense network regimes. Next, to overcome the computational intractability of the ML method, we propose a simple closed-form estimate obtained from a one-step approximation to the likelihood equation. We show that this estimate attains the same asymptotic distribution and variance as the ML estimate, thereby yielding a computationally efficient and asymptotically valid confidence interval for the natural parameter. We complement these inferential results by establishing a H\'ajek--Le Cam-type local asymptotic minimax theorem, showing that the proposed estimate achieves the smallest possible asymptotic maximum risk, both in rate and in leading constant, over shrinking neighborhoods of the true parameter. We also derive the corresponding limit of experiments. To the best of our knowledge, these are among the first sharp asymptotic optimality results for network-dependent data. Finally, we study goodness-of-fit testing for the natural parameter, deriving the local power of the likelihood ratio test and minimax detection rates. Our analysis relies on new fluctuation results for the sufficient statistic (Hamiltonian) and for the random partition function of Ising models on inhomogeneous random graphs, which are of independent interest.

Figures

Figures reproduced from arXiv: 2606.07065 by the authors.

Figure 1
Figure 1. Histograms of pβ˜N ´ βq{? θN when GN is generated from (a) an Erd˝os-R´enyi model, (b) the graphon Wpx, yq “ sinp|x´y|q, (c) W as in (2.10) with p “ 0 and q “ 1 (which corresponds to a random bipartite graph), and (d) the rank one graphon Wpx, yq “ xy. The red curves represent the limiting normal densities from Theorem 2.2. in red. The empirical coverage in the four cases are 95%, 92%, 93% and 98%, respectively, whi… view at source ↗
Figure 2
Figure 2. 100 instances of 95% confidence intervals, where GN is generated from (a) an Erd˝os-R´enyi model, (b) the graphon Wpx, yq “ sinp|x ´ y|q, (c) a random bipartite graph, and (d) the rank one graphon Wpx, yq “ xy. Here, N “ 500 and θN “ N ´0.6 . under appropriate regularity conditions, the ML estimate based on N i.i.d. samples is locally asymptotically minimax, that is, it achieves the smallest possible asymptotic maxi… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

84 extracted references · 10 canonical work pages

  1. [1]

    Animashree Anandkumar, Vincent Y. F. Tan, Furong Huang, and Allan S. Willsky. High-dimensional structure estimation in Ising models: Local separation criterion.The Annals of Statistics, 40(3):1346– 1375, 2012

    2012

  2. [2]

    Fine asymptotics of the magnetization of the annealed dilute Curie-Weiss model.arXiv:2603.09672, 2026

    Fabian Apostel, Hanna D¨ oring, and Kristina Schubert. Fine asymptotics of the magnetization of the annealed dilute Curie-Weiss model.arXiv:2603.09672, 2026

    work page arXiv 2026

  3. [3]

    Carlin, and Alan E

    Sudipto Banerjee, Bradley P. Carlin, and Alan E. Gelfand.Hierarchical Modeling and Analysis for Spatial Data. Chapman & Hall/CRC Monographs on Statistics and Applied Probability. Chapman and Hall/CRC, Boca Raton, 2nd edition, 2014

    2014

  4. [4]

    Universality of the mean-field for the Potts model.Probability Theory and Related Fields, 168(3):557–600, 2017

    Anirban Basak and Sumit Mukherjee. Universality of the mean-field for the Potts model.Probability Theory and Related Fields, 168(3):557–600, 2017

    2017

  5. [5]

    Exact recovery in the Ising blockmodel.The Annals of Statistics, 47(4):1805–1834, August 2019

    Quentin Berthet, Philippe Rigollet, and Piyush Srivastava. Exact recovery in the Ising blockmodel.The Annals of Statistics, 47(4):1805–1834, August 2019

    2019

  6. [6]

    Spatial interaction and the statistical analysis of lattice systems.Journal of the Royal Statistical Society

    Julian Besag. Spatial interaction and the statistical analysis of lattice systems.Journal of the Royal Statistical Society. Series B (Methodological), 36:192–236, 1974

    1974

  7. [7]

    Statistical analysis of non-lattice data.The Statistician, 24(3):179–195, 1975

    Julian Besag. Statistical analysis of non-lattice data.The Statistician, 24(3):179–195, 1975

    1975

  8. [8]

    Universal limit theorems in graph coloring problems with connections to extremal combinatorics.The Annals of Applied Probability, 27(1):337–394, 2017

    Bhaswar B Bhattacharya, Persi Diaconis, and Sumit Mukherjee. Universal limit theorems in graph coloring problems with connections to extremal combinatorics.The Annals of Applied Probability, 27(1):337–394, 2017

    2017

Show all 84 references
  1. [9]

    Inference in Ising models.Bernoulli, 24(1):493–525, 2018

    Bhaswar B Bhattacharya and Sumit Mukherjee. Inference in Ising models.Bernoulli, 24(1):493–525, 2018

    2018

  2. [10]

    The Ising model on a two-community stochastic block model.arXiv:2604.20631, 2026

    Alessandra Bianchi, Vanessa Jacquier, and Matteo Sfragara. The Ising model on a two-community stochastic block model.arXiv:2604.20631, 2026. ISING MODELS ON INHOMOGENEOUS RANDOM GRAPHS 25

    2026 arXiv

  3. [11]

    A nonparametric view of network models and Newman–Girvan and other modularities.Proceedings of the National Academy of Sciences, 106(50):21068–21073, 2009

    Peter J Bickel and Aiyou Chen. A nonparametric view of network models and Newman–Girvan and other modularities.Proceedings of the National Academy of Sciences, 106(50):21068–21073, 2009

    2009

  4. [12]

    Bickel, Aiyou Chen, and Elizaveta Levina

    Peter J. Bickel, Aiyou Chen, and Elizaveta Levina. The method of moments and degree distributions for network models.The Annals of Statistics, 39(5):2280–2301, 2011

    2011

  5. [13]

    The phase transition in inhomogeneous random graphs.Random Structures & Algorithms, 31(1):3–122, 2007

    B´ ela Bollob´ as, Svante Janson, and Oliver Riordan. The phase transition in inhomogeneous random graphs.Random Structures & Algorithms, 31(1):3–122, 2007

    2007

  6. [14]

    J. A. Bondy and U. S. R. Murty.Graph Theory, volume 244 ofGraduate Texts in Mathematics. Springer, New York, 2008

    2008

  7. [15]

    Chayes, L´ aszl´ o Lov´ asz, Vera T

    Christian Borgs, Jennifer T. Chayes, L´ aszl´ o Lov´ asz, Vera T. S´ os, and Katalin Vesztergombi. Conver- gent sequences of dense graphs I: Subgraph frequencies, metric properties and testing.Advances in Mathematics, 219(6):1801–1851, 2008

    2008

  8. [16]

    Chayes, L´ aszl´ o Lov´ asz, Vera T

    Christian Borgs, Jennifer T. Chayes, L´ aszl´ o Lov´ asz, Vera T. S´ os, and Katalin Vesztergombi. Convergent sequences of dense graphs II. multiway cuts and statistical physics.Annals of Mathematics, 176(1):151– 219, 2012

    2012

  9. [17]

    The thermodynamics of the Curie–Weiss model with random couplings.Journal of Statistical Physics, 72(3-4):643–664, 1993

    Anton Bovier and V´ eronique Gayrard. The thermodynamics of the Curie–Weiss model with random couplings.Journal of Statistical Physics, 72(3-4):643–664, 1993

    1993

  10. [18]

    Efficiently learning Ising models on arbitrary graphs

    Guy Bresler. Efficiently learning Ising models on arbitrary graphs. InProceedings of the 47th Annual ACM Symposium on Theory of Computing (STOC), pages 771–782. ACM, 2015

    2015

  11. [19]

    High temperature structure detection in ferromagnets.Infor- mation and Inference: A Journal of the IMA, 11(1):55–102, 2022

    Yuan Cao, Matey Neykov, and Han Liu. High temperature structure detection in ferromagnets.Infor- mation and Inference: A Journal of the IMA, 11(1):55–102, 2022

    2022

  12. [20]

    Estimation in spin glasses: A first step.The Annals of Statistics, 35(5):1931–1946, 2007

    Sourav Chatterjee. Estimation in spin glasses: A first step.The Annals of Statistics, 35(5):1931–1946, 2007

    1931

  13. [21]

    Random graphs with a given degree sequence.The Annals of Applied Probability, 21(4):1400–1435, 2011

    Sourav Chatterjee, Persi Diaconis, and Allan Sly. Random graphs with a given degree sequence.The Annals of Applied Probability, 21(4):1400–1435, 2011

    2011

  14. [22]

    Joint parameter estimations for spin glasses

    Wei-Kuo Chen, Arnab Sen, and Qiang Wu. Joint parameter estimations for spin glasses. arXiv:2406.10760, 2024

    2024

  15. [23]

    Connected components in random graphs with given expected degree sequences.Annals of Combinatorics, 6:125–145, 2002

    Fan Chung and Linyuan Lu. Connected components in random graphs with given expected degree sequences.Annals of Combinatorics, 6:125–145, 2002

    2002

  16. [24]

    Fluc- tuations of the Ising free energy on Erd˝ os–R´ enyi graphs.arXiv:2601.08590, 2026

    Amin Coja-Oghlan, Dominik Kaaser, Maurice Rolvien, Pavel Zakharov, and Kostas Zampetakis. Fluc- tuations of the Ising free energy on Erd˝ os–R´ enyi graphs.arXiv:2601.08590, 2026

    2026

  17. [25]

    On consistency of a class of estimators for exponential families of Markov random fields on the lattice.The Annals of Statistics, 20(1):455–468, 1992

    Francis Comets. On consistency of a class of estimators for exponential families of Markov random fields on the lattice.The Annals of Statistics, 20(1):455–468, 1992

    1992

  18. [26]

    Asymptotics of Maximum Likelihood Estimators for the Curie-Weiss Model.The Annals of Statistics, 19(2):557–578, 1991

    Francis Comets and Basilis Gidas. Asymptotics of Maximum Likelihood Estimators for the Curie-Weiss Model.The Annals of Statistics, 19(2):557–578, 1991

    1991

  19. [27]

    CRC Press, 2018

    Harry Crane.Probabilistic Foundations of Statistical Network Analysis. CRC Press, 2018

    2018

  20. [28]

    Learning Ising models from one or multiple samples

    Yuval Dagan, Constantinos Daskalakis, Nishanth Dikkala, and Anthimos Vardis Kandiros. Learning Ising models from one or multiple samples. InProceedings of the 53rd Annual ACM SIGACT Symposium on Theory of Computing, pages 161–168, 2021

    2021

  21. [29]

    Testing Ising models.IEEE Trans- actions on Information Theory, 65(11):6829–6852, 2019

    Constantinos Daskalakis, Nishanth Dikkala, and Gautam Kamath. Testing Ising models.IEEE Trans- actions on Information Theory, 65(11):6829–6852, 2019

    2019

  22. [30]

    Regression from dependent observa- tions

    Constantinos Daskalakis, Nishanth Dikkala, and Ioannis Panageas. Regression from dependent observa- tions. InProceedings of the 51st Annual ACM SIGACT Symposium on Theory of Computing (STOC), pages 881–889, 2019

    2019

  23. [31]

    Logistic regression with peer-group effects via inference in higher-order Ising models

    Constantinos Daskalakis, Nishanth Dikkala, and Ioannis Panageas. Logistic regression with peer-group effects via inference in higher-order Ising models. InProceedings of the 23rd International Conference on Artificial Intelligence and Statistics (AISTATS), pages 3653–3663, 2020

    2020

  24. [32]

    Pivotal CLTs for pseudolikelihood via conditional centering in dependent random fields

    Nabarun Deb. Pivotal CLTs for pseudolikelihood via conditional centering in dependent random fields. arXiv:2510.04972, 2025

    2025

  25. [33]

    Detecting structured signals in Ising models.The Annals of Applied Probability, 34(1A):1–45, 2024

    Nabarun Deb, Rajarshi Mukherjee, Sumit Mukherjee, and Ming Yuan. Detecting structured signals in Ising models.The Annals of Applied Probability, 34(1A):1–45, 2024. 26 MUKHERJEE, BHOWAL, CHATTERJEE, AND BHATTACHARYA

    2024

  26. [34]

    Fluctuations in mean-field Ising models.The Annals of Applied Probability, 33(3):1961–2003, 2023

    Nabarun Deb and Sumit Mukherjee. Fluctuations in mean-field Ising models.The Annals of Applied Probability, 33(3):1961–2003, 2023

    1961

  27. [35]

    Ising model of financial markets with many assets.Physica A: Statistical Mechanics and its Applications, 462:250–254, 2016

    A Eckrot, J Jurczyk, and I Morgenstern. Ising model of financial markets with many assets.Physica A: Statistical Mechanics and its Applications, 462:250–254, 2016

    2016

  28. [36]

    Ellis.Entropy, Large Deviations, and Statistical Mechanics

    Richard S. Ellis.Entropy, Large Deviations, and Statistical Mechanics. Springer, New York, 1985

    1985

  29. [37]

    Ellis, Charles M

    Richard S. Ellis, Charles M. Newman, and Jay S. Rosen. Limit theorems for sums of dependent random variables occurring in statistical mechanics.Zeitschrift f¨ ur Wahrscheinlichkeitstheorie und Verwandte Gebiete, 51:153–169, 1980

    1980

  30. [38]

    Ellis and Kongming Wang

    Richard S. Ellis and Kongming Wang. Limit theorems for maximum likelihood estimators in the Curie– Weiss–Potts model.Stochastic Processes and their Applications, 40(2):251–288, 1992

    1992

  31. [39]

    Markov random field image models and their applications to computer vision

    Stuart Geman and Christine Graffigne. Markov random field image models and their applications to computer vision. InProceedings of the International Congress of Mathematicians, pages 1496–1517, Berkeley, California, 1986

    1986

  32. [40]

    Joint estimation of parameters in Ising model.The Annals of Statistics, 48(2):785–810, 2020

    Promit Ghosal and Sumit Mukherjee. Joint estimation of parameters in Ising model.The Annals of Statistics, 48(2):785–810, 2020

    2020

  33. [41]

    Consistency of maximum likelihood and pseudolikelihood estimators for Gibbs distri- butions

    Basilis Gidas. Consistency of maximum likelihood and pseudolikelihood estimators for Gibbs distri- butions. In W. Fleming and P.-L. Lions, editors,Stochastic Differential Systems, Stochastic Control Theory and Applications, pages 129–145. Springer, New York, 1988

    1988

  34. [42]

    Green and Sylvia Richardson

    Peter J. Green and Sylvia Richardson. Hidden Markov models and disease mapping.Journal of the American Statistical Association, 97(460):1055–1070, 2002

    2002

  35. [43]

    Springer, 2006

    Allan Gut.Probability: A Graduate Course, volume 200. Springer, 2006

    2006

  36. [44]

    Information theoretic properties of Markov Random Fields, and their algorithmic applications

    Linus Hamilton, Frederic Koehler, and Ankur Moitra. Information theoretic properties of Markov Random Fields, and their algorithmic applications. InAdvances in Neural Information Processing Systems (NeurIPS), pages 2463–2472, 2017

    2017

  37. [45]

    Latent space approaches to social network analysis.Journal of The American Statistical Association, 97(460):1090–1098, 2002

    Peter D Hoff, Adrian E Raftery, and Mark S Handcock. Latent space approaches to social network analysis.Journal of The American Statistical Association, 97(460):1090–1098, 2002

    2002

  38. [46]

    Stochastic blockmodels: First steps.Social Networks, 5(2):109–137, 1983

    Paul W Holland, Kathryn Blackmond Laskey, and Samuel Leinhardt. Stochastic blockmodels: First steps.Social Networks, 5(2):109–137, 1983

    1983

  39. [47]

    Hopfield

    John J. Hopfield. Neural networks and physical systems with emergent collective computational abilities. Proceedings of the National Academy of Sciences of the United States of America, 79(8):2554–2558, 1982

    1982

  40. [48]

    Beitrag zur theorie des ferromagnetismus.Zeitschrift f¨ ur Physik, 31:253–258, 1925

    Ernst Ising. Beitrag zur theorie des ferromagnetismus.Zeitschrift f¨ ur Physik, 31:253–258, 1925

    1925

  41. [49]

    Fluctuations of the magnetization for Ising models on dense Erd˝ os–R´ enyi random graphs.Journal of Statistical Physics, 177:78–94, 2019

    Zakhar Kabluchko, Matthias L¨ owe, and Kristina Schubert. Fluctuations of the magnetization for Ising models on dense Erd˝ os–R´ enyi random graphs.Journal of Statistical Physics, 177:78–94, 2019

    2019

  42. [50]

    Zakhar Kabluchko, Matthias L¨ owe, and Kristina Schubert. Fluctuations of the magnetization for Ising models on Erd˝ os–R´ enyi random graphs–the regimes of smallpand the critical temperature.Journal of Physics A: Mathematical and Theoretical, 53(35):355004, 2020

    2020

  43. [51]

    Zakhar Kabluchko, Matthias L¨ owe, and Kristina Schubert. Fluctuations for the partition function of Ising models on Erd˝ os–R´ enyi random graphs.Annales de l’Institut Henri Poincare (B) Probabilites et statistiques, 57(4):2017–2042, 2021

    2017

  44. [52]

    Fluctuations of the magnetization for Ising models on Erd˝ os–R´ enyi random graphs – the regimes of low temperature and external magnetic field

    Zakhar Kabluchko, Matthias L¨ owe, and Kristina Schubert. Fluctuations of the magnetization for Ising models on Erd˝ os–R´ enyi random graphs – the regimes of low temperature and external magnetic field. ALEA– Latin American Journal of Probability and Mathematical Statistics, ...

    2022

  45. [53]

    Statistically valid variational Bayes al- gorithm for ising model parameter estimation.Journal of Computational and Graphical Statistics, 33(1):75–84, 2024

    Minwoo Kim, Shrijita Bhattacharya, and Tapabrata Maiti. Statistically valid variational Bayes al- gorithm for ising model parameter estimation.Journal of Computational and Graphical Statistics, 33(1):75–84, 2024

    2024

  46. [54]

    Learning graphical models using multiplicative weights

    Adam Klivans and Raghu Meka. Learning graphical models using multiplicative weights. InProceedings of the 58th IEEE Annual Symposium on Foundations of Computer Science (FOCS), pages 343–354. IEEE, 2017

    2017

  47. [55]

    Limits of experiments

    Lucien Le Cam. Limits of experiments. InProceedings of the Sixth Berkeley Symposium on Mathematical Statistics and Probability, volume 1, pages 245–261, Berkeley, Los Angeles, 1972. University of California Press. ISING MODELS ON INHOMOGENEOUS RANDOM GRAPHS 27

    1972

  48. [56]

    CLT in high-dimensional Bayesian linear regres- sion with low SNR.arXiv:2507.23285, 2025

    Seunghyun Lee, Nabarun Deb, and Sumit Mukherjee. CLT in high-dimensional Bayesian linear regres- sion with low SNR.arXiv:2507.23285, 2025

    2025

  49. [57]

    Fluctuations in random field Ising models

    Seunghyun Lee, Nabarun Deb, and Sumit Mukherjee. Fluctuations in random field Ising models. arXiv:2503.21152, 2025

    2025 arXiv

  50. [58]

    American Mathematical Soc., 2012

    L´ aszl´ o Lov´ asz.Large networks and graph limits, volume 60. American Mathematical Soc., 2012

    2012

  51. [59]

    Inference in high-dimensional logistic regression under tensor network dependence.arXiv:2603.20082, 2026

    Josh Miles and Sohom Bhattacharya. Inference in high-dimensional logistic regression under tensor network dependence.arXiv:2603.20082, 2026

    2026

  52. [60]

    The spread of innovations in social networks.Proceedings of the National Academy of Sciences, 107(47):20196–20201, 2010

    Andrea Montanari and Amin Saberi. The spread of innovations in social networks.Proceedings of the National Academy of Sciences, 107(47):20196–20201, 2010

    2010

  53. [61]

    Global testing against sparse alternatives under Ising models.The Annals of Statistics, 46(5):2062–2093, 2018

    Rajarshi Mukherjee, Sumit Mukherjee, and Ming Yuan. Global testing against sparse alternatives under Ising models.The Annals of Statistics, 46(5):2062–2093, 2018

  54. [62]

    On testing for parameters in Ising models.Annales de l’Institut Henri Poincar´ e – Probabilit´ es et Statistiques, 58(1):164–187, 2022

    Rajarshi Mukherjee and Gourab Ray. On testing for parameters in Ising models.Annales de l’Institut Henri Poincar´ e – Probabilit´ es et Statistiques, 58(1):164–187, 2022

    2022

  55. [63]

    Bhattacharya, and George Michailidis

    Somabha Mukherjee, Ziang Niu, Sagnik Halder, Bhaswar B. Bhattacharya, and George Michailidis. High dimensional logistic regression under network dependence.Journal of Machine Learning Research, 25:1–62, 2024

    2024

  56. [64]

    Property testing in high-dimensional Ising models.The Annals of Statistics, 47(5):2472–2503, 2019

    Matey Neykov and Han Liu. Property testing in high-dimensional Ising models.The Annals of Statistics, 47(5):2472–2503, 2019

    2019

  57. [65]

    Asymptotic distributions of weightedU-statistics of degree 2

    Kevin A O’Neil and Richard A Redner. Asymptotic distributions of weightedU-statistics of degree 2. The Annals of Probability, 21(2):1159–1169, 1993

    1993

  58. [66]

    David K. Pickard. Asymptotic inference for an ising lattice.Journal of Applied Probability, 13(3):486– 497, 1976

    1976

  59. [67]

    David K. Pickard. Asymptotic inference for an ising lattice. II.Advances in Applied Probability, 9(3):476– 501, 1977

    1977

  60. [68]

    David K. Pickard. Asymptotic inference for an ising lattice. III. non-zero field and ferromagnetic states. Journal of Applied Probability, 16(1):12–24, 1979

    1979

  61. [69]

    David K. Pickard. Inference for discrete Markov fields: The simplest nontrivial case.Journal of the American Statistical Association, 82(397):90–96, 1987

    1987

  62. [70]

    Distribution of the magnetization of the critical Ising model on sparse random graphs.arXiv:2603.28702, 2026

    Kyprianos-Iason Prodromidis and Allan Sly. Distribution of the magnetization of the critical Ising model on sparse random graphs.arXiv:2603.28702, 2026

    2026

  63. [71]

    A polynomial-time algorithm for image segmentation using Ising models

    Peiwei Qin and Jieyu Zhao. A polynomial-time algorithm for image segmentation using Ising models. In2011 Seventh International Conference on Natural Computation, volume 2, pages 932–935, 2011

    2011

  64. [72]

    Wainwright, and John D

    Pradeep Ravikumar, Martin J. Wainwright, and John D. Lafferty. High-dimensional Ising model selec- tion usingℓ 1-regularized logistic regression.The Annals of Statistics, 38(3):1287–1319, 2010

    2010

  65. [73]

    Hanson-wright inequality and sub-gaussian concentration.Elec- tronic Communications in Probability, 18:1–9, 2013

    Mark Rudelson and Roman Vershynin. Hanson-wright inequality and sub-gaussian concentration.Elec- tronic Communications in Probability, 18:1–9, 2013

    2013

  66. [74]

    Santhanam and Martin J

    Narayana P. Santhanam and Martin J. Wainwright. Information-theoretic limits of selecting binary graphical models in high dimensions.IEEE Transactions on Information Theory, 58(7):4117–4134, 2012

    2012

  67. [75]

    Stoˇ si´ c and Ivon P

    Borko D. Stoˇ si´ c and Ivon P. Fittipaldi. Pattern recognition via Ising model with long range interactions. Physica A: Statistical Mechanics and its Applications, 242(3-4):323–331, 1997

    1997

  68. [76]

    Cambridge university press, 2000

    Aad W Van der Vaart.Asymptotic statistics, volume 3. Cambridge university press, 2000

    2000

  69. [77]

    van der Vaart and Jon A

    Aad W. van der Vaart and Jon A. Wellner.Weak Convergence and Empirical Processes: With Appli- cations to Statistics. Springer, New York, 1996

    1996

  70. [78]

    Cambridge university press, 2018

    Roman Vershynin.High-dimensional probability: An introduction with applications in data science, volume 47. Cambridge university press, 2018

    2018

  71. [79]

    Lokhov, and Michael Chertkov

    Marc Vuffray, Sidhant Misra, Andrey Y. Lokhov, and Michael Chertkov. Interaction screening: Efficient and sample-optimal learning of Ising models.arXiv:1605.07252, 2016

    2016 arXiv

  72. [80]

    1. Then for anypě1, eγϑN `γ 2ηN “1`γ ϑN `γ 2ηN `o Lp ˆ 1?N θN ˙ .(A.7) Furthermore, ifN θ 2 N

    Yuanzhe Xu and Sumit Mukherjee. Inference in Ising models on dense regular graphs.The Annals of Statistics, 51(3):1183–1206, 2023. 28 MUKHERJEE, BHOWAL, CHATTERJEE, AND BHATTACHARYA AppendixA.Proof of Theorem 5.2 Recall the definition of the partition functionZ N of the model ...

    2023

  73. [81]

    ˜XN ` ¯XN, Lemma A.8 implies thatE µN |UN |1`δ ≲1, for someδą0. Therefore, by H¨ older’s inequality, EµN |VN | ď ´ EµN |UN |1`δ ¯ 1 1`δ ˜ EµN

    Hence, there exists a constantL 0 ą0 such that, N θN EµN “ RN pσ, τqETpσ qETpτ q ‰ ≲β N θN EµN « RN pσ, τqe β N ř 1ďiăjďN W ´ i N , j N ¯ pσiσj `τ iτjq `L 0 pQN pσq `Q N pτ qq ff “E µN rUN s `E µN rVN s,(A.45) where UN :“N θ N RN pσ, τqe β N ř 1ďiăjďN W ´ i N , j N ¯ pσiσj `τ ...

  74. [82]

    P ˜ N ˜ 0, 2´ r0,1s2 Wpx, yqdxdy ¸ ďt ¸ . This completes the proof of Theorem 2.1 (1).2 C.1.2.Proof of Theorem 2.1p2q.Recall from the proof of Theorem 5.1, that for everytPR, Eβ

    This proves (5.2). B.2.Proof of Theorem 5.1 (2).In this case, e t 2N řN i“1 W ` i N , i N ˘ ` t2`2βt 4N2θN ř 1ďi,jďN W ´ i N , j N ¯ Ñe t2`2tβ 4θ ´ r0,1s2 Wpx, yqdxdy` t 2 ´ 1 0 Wpx, xqdx .(B.9) We also have the following from Lemma A.4: E ˆZN pβ`tq E ˆZN pβq Ñe ´ t2`2tβ 4 ´ r...

  75. [83]

    e řN i“1 f p1q i ` ř 1ďiăjďN f p2q ij pσiσjq.(F.5) Expanding the exponential and the logarithm inf 1 i and using the factN θ N

    Then we have from (D.25) lim sup NÑ8 P ˜ sup β1,N PΘpβ0,δN q Pβ1,N pϕα N “0|A GN q ąε ¸ ďlim sup NÑ8 P pmaxtAN,M , BN,M u ąε q ďlim sup NÑ8 P ´ maxtAN,M , BN,M u ´c M ą ε 2 ¯ “0, thereby showing that sup β1,N PΘpβ0,δN q Pβ1,N pϕα N “0|A GN q PÑ0. Combining this with (D.23) sho...

  76. [84]

    1, such that E

    Indeed, in the following we will show, }WN ´W} 1 Ñ0 ISING MODELS ON INHOMOGENEOUS RANDOM GRAPHS 79 which will complete the proof by using the inequalityδ 2 pWN , Wq ď }W N ´W} 1, where} ¨ } 1 is theL 1 norm (see [58, Chapter 8]). Towards that define, UN pWq:“ 1 N 2 ÿ 1ďu,vďN s...

Pith tools

Reviewed June 27, 2026 · model on record in the stance chip above.