Asymptotic analysis of the stochastic block model for modular networks and its algorithmic applications

@article{Decelle2011AsymptoticAO,
  title={Asymptotic analysis of the stochastic block model for modular networks and its algorithmic applications},
  author={Aur{\'e}lien Decelle and Florent Krzakala and Cristopher Moore and Lenka Zdeborov{\'a}},
  journal={Physical review. E, Statistical, nonlinear, and soft matter physics},
  year={2011},
  volume={84 6 Pt 2},
  pages={
          066106
        }
}
In this paper we extend our previous work on the stochastic block model, a commonly used generative model for social and biological networks, and the problem of inferring functional groups or communities from the topology of the network. We use the cavity method of statistical physics to obtain an asymptotically exact analysis of the phase diagram. We describe in detail properties of the detectability-undetectability phase transition and the easy-hard phase transition for the community… 

Figures from this paper

A Regularized Stochastic Block Model for the robust community detection in complex networks

This work shows experimentally that given the regularized model, the inference algorithms, such as Markov chain Monte Carlo, reliably and quickly find the assortative or disassortative structure as directed by the value of a single parameter.

Detectability thresholds and optimal algorithms for community structure in dynamic networks

The fundamental limits on learning latent community structure in dynamic networks where nodes change their community membership over time, but where edges are generated independently at each time step are studied, and it is claimed that no algorithm can identify the communities better than chance.

A Fast Inference Algorithm for Stochastic Blockmodel

This paper proposes a fast and principled inference algorithm for stochastic block model, based on the variational Bayesian framework, and deploys the natural conjugate gradient method to accelerate the optimization of the Variational bound.

Large deviations of connected components in the stochastic block model.

The stochastic block model is studied, which is often used to model community structures and study community-detection algorithms, and the distributions of their sizes including the tails down to probabilities smaller than 10^{-800}.

Phase transitions and optimal algorithms for semisupervised classifications on graphs: From belief propagation to graph convolution network

For the first time, well-controlled benchmark datasets with asymptotically exact properties and optimal solutions could be produced for the evaluation of graph convolution neural networks, and for the theoretical understanding of their strengths and weaknesses.

An improved belief propagation algorithm for detecting meso-scale structure in complex networks

The original and the improved BP algorithms yield a similar performance regarding the community detection; however, the improvedBP algorithm is much better and more stable when the CP structure becomes more dominant, which may help us correctly partition different types of mesoscale structures in networks.

Mutual information for the sparse stochastic block model

A conjecture for the limit of this quantity is expressed in terms of a Hamilton-Jacobi equation posed over a space of probability measures, and a proof that this conjectured limit provides a lower bound for the asymptotic mutual information is shown.

Optimal rates for community estimation in the weighted stochastic block model

A weighted generalization of the stochastic block models, in which observations are collected in the form of a weighted adjacency matrix and the weight of each edge is generated independently from an unknown probability density determined by the community membership of its endpoints, is studied.

Relative Density and Exact Recovery in Heterogeneous Stochastic Block Models

It is shown that it is possible, in the right circumstances, to recover very small clusters (up to $\sqrt{\log n}$ size), if there are just a few of them (at most polylogarithmic in $n$).

Disentangling group and link persistence in dynamic stochastic block models

It is analytically and numerically shown that communities inferred from single network snapshot can share a maximum overlap with the underlying communities of a specific previous instant in time, which leads to time-lagged inference: the identification of past communities rather than present ones.
...

References

SHOWING 1-10 OF 114 REFERENCES

Phase transition in the detection of modules in sparse networks

An asymptotically exact analysis of the problem of detecting communities in sparse random networks generated by stochastic block models using the cavity method of statistical physics and its relationship to belief propagation yields an optimal inference algorithm for detecting modules.

Mixture models and exploratory analysis in networks

A general technique for detecting structural features in large-scale network data that works by dividing the nodes of a network into classes such that the members of each class have similar patterns of connection to other nodes is described.

Stochastic blockmodels and community structure in networks

  • B. KarrerM. Newman
  • Computer Science
    Physical review. E, Statistical, nonlinear, and soft matter physics
  • 2011
This work demonstrates how the generalization of blockmodels to incorporate this missing element leads to an improved objective function for community detection in complex networks and proposes a heuristic algorithm forcommunity detection using this objective function or its non-degree-corrected counterpart.

An efficient and principled method for detecting communities in networks

This work describes a method for finding overlapping communities based on a principled statistical approach using generative network models and shows how the method can be implemented using a fast, closed-form expectation-maximization algorithm that allows us to analyze networks of millions of nodes in reasonable running times.

Estimation and Prediction for Stochastic Blockstructures

A statistical approach to a posteriori blockmodeling for digraphs and valued digraphs is proposed. The probability model assumes that the vertices of the digraph are partitioned into several

A nonparametric view of network models and Newman–Girvan and other modularities

An attempt at unifying points of view and analyses of these objects coming from the social sciences, statistics, probability and physics communities are presented and the approach to the Newman–Girvan modularity, widely used for “community” detection, is applied.

Benchmark graphs for testing community detection algorithms.

This work introduces a class of benchmark graphs, that account for the heterogeneity in the distributions of node degrees and of community sizes, and uses this benchmark to test two popular methods of community detection, modularity optimization, and Potts model clustering.

Missing and spurious interactions and the reconstruction of complex networks

This work is able to reliably identify both missing and spurious interactions in noisy network observations and enables network reconstructions that yield estimates of the true network properties that are more accurate than those provided by the observations themselves.

(Un)detectable cluster structure in sparse networks.

A sharp transition is found from a phase in which the cluster structure is not detectable at all to aphase in which it can be detected with high accuracy, which illuminates theoretical limitations of data mining in networks and allows for an understanding of the performance of a variety of algorithms.

Community detection in graphs

...