Pair Matching: When bandits meet stochastic block model

The pair-matching problem appears in many applications where one wants to discover good matches between pairs of individuals. Formally, the set of individuals is represented by the nodes of a graph where the edges, unobserved at first, represent the good matches. The algorithm queries pairs of nodes and observes the presence/absence of edges. Its goal is to discover as many edges as possible with a fixed budget of queries. Pair-matching is a particular instance of multi-armed bandit problem in which the arms are pairs of individuals and the rewards are edges linking these pairs. This bandit problem is non-standard though, as each arm can only be played once. Given this last constraint, sublinear regret can be expected only if the graph presents some underlying structure. This paper shows that sublinear regret is achievable in the case where the graph is generated according to a Stochastic Block Model (SBM) with two communities. Optimal regret bounds are computed for this pair-matching problem. They exhibit a phase transition related to the Kesten-Stigund threshold for community detection in SBM. To avoid undesirable features of optimal solutions, the pair-matching problem is also considered in the case where each node is constrained to be sampled less than a given amount of times. We show how this constraint deteriorates optimal regret rates. The paper is concluded by a conjecture regarding the optimal regret when the number of communities is larger than $2$. Contrary to the two communities case, we believe that a statistical-computational gap would appear in this problem.

[1]  P. Massart,et al.  Concentration inequalities and model selection , 2007 .

[2]  Alexandre Proutière,et al.  Accurate Community Detection in the Stochastic Block Model via Spectral Algorithms , 2014, ArXiv.

[3]  Yu Lu,et al.  Statistical and Computational Guarantees of Lloyd's Algorithm and its Variants , 2016, ArXiv.

[4]  H. Robbins.  Some aspects of the sequential design of experiments , 1952 .

[5]  Kathryn B. Laskey,et al.  Stochastic blockmodels: First steps , 1983 .

[6]  M. Lelarge,et al.  Nonbacktracking spectrum of random graphs: Community detection and nonregular Ramanujan graphs , 2018 .

[7]  Minje Jang,et al.  Top-K Ranking from Pairwise Comparisons: When Spectral Ranking is Optimal , 2016, ArXiv.

[8]  A. Burnetas,et al.  Optimal Adaptive Policies for Sequential Allocation Problems , 1996 .

[9]  David Firth,et al.  Ultraviolet signals ultra-aggression in a lizard , 2006, Animal Behaviour.

[10]  Elchanan Mossel,et al.  Consistency thresholds for the planted bisection model , 2016 .

[11]  Zhengxing Chen,et al.  EOMM: An Engagement Optimized Matchmaking Framework , 2017, WWW.

[12]  Jess Banks,et al.  Information-theoretic thresholds for community detection in sparse networks , 2016, COLT.

[13]  Chao Gao,et al.  Achieving Optimal Misclassification Proportion in Stochastic Block Models , 2015, J. Mach. Learn. Res..

[14]  S. Redner,et al.  Understanding baseball team standings and streaks , 2008 .

[15]  Emmanuel Abbe,et al.  Community detection and stochastic block models: recent developments , 2017, Found. Trends Commun. Inf. Theory.

[16]  Martin J. Wainwright,et al.  Simple, Robust and Optimal Ranking from Pairwise Comparisons , 2015, J. Mach. Learn. Res..

[17]  Harry Joe,et al.  Extended Use of Paired Comparison Models, with Application to Chess Rankings , 1990 .

[18]  Emmanuel Abbe,et al.  Detection in the stochastic block model with multiple clusters: proof of the achievability conjectures, acyclic BP, and the information-computation gap , 2015, ArXiv.

[19]  Yudong Chen,et al.  Exponential Error Rates of SDP for Block Models: Beyond Grothendieck’s Inequality , 2017, IEEE Transactions on Information Theory.

[20]  R. A. Bradley,et al.  RANK ANALYSIS OF INCOMPLETE BLOCK DESIGNS , 1952 .

[21]  E. Zermelo.  Die Berechnung der Turnier-Ergebnisse als ein Maximumproblem der Wahrscheinlichkeitsrechnung , 1929 .

[22]  Emmanuel Abbe,et al.  Community Detection in General Stochastic Block models: Fundamental Limits and Efficient Algorithms for Recovery , 2015, 2015 IEEE 56th Annual Symposium on Foundations of Computer Science.

[23]  Christophe Giraud,et al.  Partial recovery bounds for clustering with the relaxed K-means , 2018, Mathematical Statistics and Learning.

[24]  Anderson Y. Zhang,et al.  Minimax Rates of Community Detection in Stochastic Block Models , 2015, ArXiv.

[25]  Sylvain Le Corff,et al.  A Bayesian nonparametric approach for generalized Bradley-Terry models in random environment , 2018, 1808.08104.

[26]  Cristopher Moore,et al.  Asymptotic analysis of the stochastic block model for modular networks and its algorithmic applications , 2011, Physical review. E, Statistical, nonlinear, and soft matter physics.

[27]  Cristopher Moore,et al.  The Computer Science and Physics of Community Detection: Landscapes, Phase Transitions, and Hardness , 2017, Bull. EATCS.

[28]  Anup Rao,et al.  Stochastic Block Model and Community Detection in Sparse Graphs: A spectral algorithm with optimal rate of recovery , 2015, COLT.

[29]  Elchanan Mossel,et al.  Reconstruction and estimation in the planted partition model , 2012, Probability Theory and Related Fields.

[30]  Yordan Zaykov,et al.  TrueSkill 2: An improved Bayesian skill rating system , 2018 .

[31]  Yudong Chen,et al.  Statistical-Computational Tradeoffs in Planted Problems and Submatrix Localization with a Growing Number of Clusters and Submatrices , 2014, J. Mach. Learn. Res..

[32]  Tom Minka,et al.  TrueSkillTM: A Bayesian Skill Rating System , 2006, NIPS.

[33]  W. R. Thompson.  ON THE LIKELIHOOD THAT ONE UNKNOWN PROBABILITY EXCEEDS ANOTHER IN VIEW OF THE EVIDENCE OF TWO SAMPLES , 1933 .

[34]  Aurélien Garivier,et al.  Explore First, Exploit Next: The True Shape of Regret in Bandit Problems , 2016, Math. Oper. Res..

[35]  Aurélien Garivier,et al.  On the Complexity of Best-Arm Identification in Multi-Armed Bandit Models , 2014, J. Mach. Learn. Res..

[36]  T. L. Lai Andherbertrobbins.  Asymptotically Efficient Adaptive Allocation Rules , 1985 .

[37]  Arnaud Doucet,et al.  Efficient Bayesian Inference for Generalized Bradley–Terry Models , 2010, 1011.1761.

[38]  Alexandre Proutière,et al.  Community Detection via Random and Adaptive Sampling , 2014, COLT.

[39]  Martin J. Wainwright,et al.  Estimation from Pairwise Comparisons: Sharp Minimax Bounds with Topology Dependence , 2015, J. Mach. Learn. Res..

[40]  Laurent Massoulié,et al.  Robustness of spectral methods for community detection , 2018, COLT.

[41]  D. Hunter.  MM algorithms for generalized Bradley-Terry models , 2003 .

[42]  Laurent Massoulié,et al.  Community detection thresholds and the weak Ramanujan property , 2013, STOC.