Abstract
We consider the problem of identifying underlying community-like structures in graphs. Toward this end, we study the stochastic block model (SBM) on k-clusters: a random model on n = km vertices, partitioned in k equal sized clusters, with edges sampled independently across clusters with probability q and within clusters with probability p, p > q. The goal is to recover the initial “hidden” partition of [n]. We study semidefinite programming (SDP)-based algorithms in this context. In the regime \(p = \frac {\alpha \log (m)}{m}\) and \(q = \frac {\beta \log (m)}{m}\), we show that a certain natural SDP-based algorithm solves the problem of exact recovery in the k-community SBM, with high probability, whenever \(\sqrt {\alpha } - \sqrt {\beta } > \sqrt {1}\), as long as \(k=o(\log n)\). This threshold is known to be the information theoretically optimal. We also study the case when \(k=\theta (\log (n))\). In this case however, we achieve recovery guarantees that no longer match the optimal condition \(\sqrt {\alpha } - \sqrt {\beta } > \sqrt {1}\), thus leaving achieving optimality for this range an open question.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
Indeed by definition any vector \(y \in \mathbb {R}_{n | k} \oplus \mathbb {1}\) can be written as \(x + \delta \frac {\mathbb {1}}{\sqrt {n}}\) for some δ and \(x \in \mathbb {R}_{n | k}\). For the purpose of proving positive definiteness, we can always divide by any positive number and can therefore consider \(\frac {y}{\|x\|}\). Also note that we can consider y or − y equivalently and hence can consider the case when δ > 0.
References
E. Abbe, C. Sandon, Community detection in general stochastic block models: fundamental limits and efficient recovery algorithms (2015). Available online at arXiv:1503.00609 [math.PR]
E. Abbe, A.S. Bandeira, G. Hall, Exact recovery in the stochastic block model (2014). Available online at arXiv:1405.3267 [cs.SI]
N. Alon, N. Kahale, A spectral technique for coloring random 3-colorable graphs. SIAM J. Comput. 26(6), 1733–1748 (1997)
N. Alon, M. Krivelevich, B. Sudakov, Finding a large hidden clique in a random graph, in Proceedings of the Ninth Annual ACM-SIAM Symposium on Discrete Algorithms, 25–27 January 1998, San Francisco, CA (1998), pp. 594–598
E. Arias-Castro, N. Verzelen, Community detection in random networks (2013). Available online at arXiv:1302.7099 [math.ST]
P. Awasthi, A.S. Bandeira, M. Charikar, R. Krishnaswamy, S. Villar, R. Ward, Relax, no need to round: integrality of clustering formulations, in 6th Innovations in Theoretical Computer Science (ITCS 2015) (2015)
A.S. Bandeira, Random Laplacian matrices and convex relaxations (2015). Available online at arXiv:1504.03987 [math.PR]
A.S. Bandeira, R.V. Handel, Sharp nonasymptotic bounds on the norm of random matrices with independent entries. Ann. Probab. 44(4), 2479–2506 (2016)
A.S. Bandeira, M. Charikar, A. Singer, A. Zhu, Multireference alignment using semidefinite programming, in 5th Innovations in Theoretical Computer Science (ITCS 2014) (2014)
A.S. Bandeira, Y. Chen, A. Singer, Non-unique games over compact groups and orientation estimation in cryo-em (2015). Available at arXiv:1505.03840 [cs.CV]
R.B. Boppana, Eigenvalues and graph bisection: an average-case analysis, in Proceedings of the 28th Annual Symposium on Foundations of Computer Science, SFCS ’87, Washington, DC (IEEE Computer Society, Washington, 1987), pp. 280–285
T.N. Bui, S. Chaudhuri, F.T. Leighton, M. Sipser, Graph bisection algorithms with good average case behavior, in 25th Annual Symposium on Foundations of Computer Science, 24–26 October 1984, West Palm Beach, FL (1984), pp. 181–192
M. Charikar, K. Makarychev, Y. Makarychev, Near-optimal algorithms for unique games, in Proceedings of the Thirty-eighth Annual ACM Symposium on Theory of Computing, STOC ’06, New York, NY (ACM, New York, 2006), pp. 205–214
Y. Chen, J. Xu, Statistical-computational tradeoffs in planted problems and submatrix localization with a growing number of clusters and submatrices (2014). Available online at arXiv:1402.1267 [stat.ML]
P. Chin, A. Rao, V. Vu, Stochastic block model and community detection in the sparse graphs: A spectral algorithm with optimal rate of recovery (2015). Available online at: arXiv:1501.05021
A. Condon, R.M. Karp, Algorithms for graph partitioning on the planted partition model. Random Struct. Algor. 18(2), 116–140 (2001)
A. Decelle, F. Krzakala, C. Moore, L. Zdeborová, Asymptotic analysis of the stochastic block model for modular networks and its algorithmic applications. Phys. Rev. E 84, 066106 (2011)
P. Erdös, A. Renyi, On random graphs. I. Publ. Math. 6, 290–297 (1959)
U. Feige, J. Kilian, Heuristics for semirandom graph problems. J. Comput. Syst. Sci. 63(4), 639–671 (2001)
A.M. Frieze, M. Jerrum, Improved approximation algorithms for max k-cut and max bisection, in Proceedings of the 4th International IPCO Conference on Integer Programming and Combinatorial Optimization (Springer-Verlag, London, 1995), pp. 1–13
B. Hajek, Y. Wu, J. Xu, Achieving exact cluster recovery threshold via semidefinite programming (2014). Available online at arXiv:1412.6156 [stat.ML]
B. Hajek, Y. Wu, J. Xu, Achieving exact cluster recovery threshold via semidefinite programming: extensions (2015). Available online at arXiv:1502.07738 [stat.ML]
R. Krauthgamer, J. Naor, R. Schwartz, Partitioning graphs into balanced components, in Proceedings of the Twentieth Annual ACM-SIAM Symposium on Discrete Algorithms, SODA ’09 (Society for Industrial and Applied Mathematics, Philadelphia, PA, 2009), pp. 942–949
K. Makarychev, Y. Makarychev, A. Vijayaraghavan, Constant factor approximation for balanced cut in the PIE model, in Symposium on Theory of Computing, STOC 2014, New York, NY, May 31–June 03 (2014), pp. 41–49
L. Massoulié, Community detection thresholds and the weak Ramanujan property, in Symposium on Theory of Computing, STOC 2014, New York, NY, May 31–June 03 (2014), pp. 694–703
F. McSherry, Spectral partitioning of random graphs, in Proceedings of the 42nd IEEE Symposium on Foundations of Computer Science, FOCS ’01 (IEEE Computer SocietyWashington, DC, 2001), p. 529
E. Mossel, J. Neeman, A. Sly, Stochastic block models and reconstruction (2012). Available online at arXiv:1202.1499
E. Mossel, J. Neeman, A. Sly, A proof of the block model threshold conjecture (2013). Available online at arXiv:1311.4115
E. Mossel, J. Neeman, A. Sly, Belief propagation, robust reconstruction and optimal recovery of block models, in Proceedings of The 27th Conference on Learning Theory, COLT 2014, Barcelona, June 13–15 (2014), pp. 356–370
E. Mossel, J. Neeman, A. Sly, Consistency thresholds for binary symmetric block models (2014). Available online at arXiv: 1407.1591
V. Vu, A simple SVD algorithm for finding hidden partitions. Available online at arXiv: 1404.3918 (2014)
S.-Y. Yun, A. Proutiere, Accurate community detection in the stochastic block model via spectral algorithms (2014). Available online at arXiv: 1412.7335
Acknowledgements
Most of the work presented in this paper was conducted while ASB was at Princeton University and partly conducted while ASB was at the Massachusetts Institute of Technology. ASB acknowledges support from AFOSR Grant No. FA9550-12-1-0317, NSF DMS-1317308, NSF DMS-1712730, and NSF DMS-1719545.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
1 Appendix
Forms of Chernoff Bounds and Hoeffding Bounds Used in the Arguments
Theorem 7 (Chernoff)
Suppose X 1…X n be independent random variables taking values in {0, 1}. Let X denote their sum and let \(\mu = \mathbb {E}[X]\) be its expectation. Then for any δ > 0 it holds that
A simplified form of the above bound is the following formula (for δ ≤ 1)
Theorem 8 (Bernstein)
Suppose X 1…X n be independent random variables taking values in [−M, M]. Let X denote their sum and let \(\mu = \mathbb {E}[X]\) be its expectation, then
Corollary 1
Suppose X 1…X n are i.i.d Bernoulli variables with parameter p. Let σ = σ(X i ) = p(1 − p); then we have that for any r ≥ 0
Proof
We have that nσ 2 = np(1 − p) and M = 1. We can now choose \(t = \alpha \sigma \sqrt {n\log (r)} + \alpha \log (r)\). This implies that \(\frac {n\sigma ^2 + t/3}{t^2} \leq \frac {1}{\log (r)}\left (1/\alpha ^2 + 1/3\alpha \right ) \leq \frac {2}{\alpha \log (r)} \) which implies from Theorem 8 that \(\mathbb {P}\left (X > \mu + \alpha \sigma \sqrt {n\log (r)}+ \alpha \log (r)\right ) \leq e^{-\frac {\alpha \log (r)}{4}}.\) □
Theorem 9 (Hoeffding)
Let X 1…X n be independent random variables. Assume that the X i are bounded in the interval [a i , b i ]. Define the empirical mean of these variables as
then
Rights and permissions
Copyright information
© 2017 Springer International Publishing AG
About this chapter
Cite this chapter
Agarwal, N., Bandeira, A.S., Koiliaris, K., Kolla, A. (2017). Multisection in the Stochastic Block Model Using Semidefinite Programming. In: Boche, H., Caire, G., Calderbank, R., März, M., Kutyniok, G., Mathar, R. (eds) Compressed Sensing and its Applications. Applied and Numerical Harmonic Analysis. Birkhäuser, Cham. https://doi.org/10.1007/978-3-319-69802-1_4
Download citation
DOI: https://doi.org/10.1007/978-3-319-69802-1_4
Published:
Publisher Name: Birkhäuser, Cham
Print ISBN: 978-3-319-69801-4
Online ISBN: 978-3-319-69802-1
eBook Packages: Mathematics and StatisticsMathematics and Statistics (R0)