When can we rank well from comparisons of \(O(n\log(n))\) non-actively chosen pairs?

Ranking from pairwise comparisons is a ubiquitous problem and has been studied in disciplines ranging from statistics to operations research and from theoretical computer science to machine learning. Here we consider a general setting where outcomes of pairwise comparisons between items i and j are drawn probabilistically by flipping a coin with unknown bias Pij , and ask under what conditions on these unknown probabilities one can learn a good ranking from comparisons of only O(n log n) non-actively chosen pairs. Recent work has established this is possible under the Bradley-Terry-Luce (BTL) and noisy permutation (NP) models. Here we introduce a broad family of ‘low-rank’ conditions on the probabilities Pij under which the resulting preference matrix P has low rank under some link function, and show these conditions encompass the BTL and Thurstone classes as special cases, but are considerably more general. We then give a new algorithm called low-rank pairwise ranking (LRPR) which provably learns a good ranking from comparisons of only O(n log n) randomly chosen comparisons under such low-rank models. Our algorithm and analysis make use of tools from the theory of low-rank matrix completion, and provide a new perspective on the problem of ranking from pairwise comparisons in non-active settings.

[1]  Arun Rajkumar,et al.  A Statistical Convergence Perspective of Algorithms for Rank Aggregation from Pairwise Data , 2014, ICML.

[2]  L. Thurstone.  A law of comparative judgment. , 1994 .

[3]  Thorsten Joachims,et al.  The K-armed Dueling Bandits Problem , 2012, COLT.

[4]  Craig Boutilier,et al.  Learning Mallows Models with Pairwise Preferences , 2011, ICML.

[5]  Mark Braverman,et al.  Sorting from Noisy Information , 2009, ArXiv.

[6]  Nir Ailon,et al.  An Active Learning Algorithm for Ranking from Pairwise Preferences with an Almost Optimal Query Complexity , 2010, J. Mach. Learn. Res..

[7]  R. A. Bradley,et al.  Rank Analysis of Incomplete Block Designs: I. The Method of Paired Comparisons , 1952 .

[8]  Noga Alon,et al.  Ranking Tournaments , 2006, SIAM J. Discret. Math..

[9]  Eyke Hüllermeier,et al.  A Survey of Preference-Based Online Learning with Bandit Algorithms , 2014, ALT.

[10]  Jean Lafond,et al.  Low Rank Matrix Completion with Exponential Family Noise , 2015, COLT.

[11]  Atri Rudra,et al.  Ordering by weighted number of wins gives a good ranking for weighted tournaments , 2006, SODA '06.

[12]  Emmanuel J. Candès,et al.  A Singular Value Thresholding Algorithm for Matrix Completion , 2008, SIAM J. Optim..

[13]  David F. Gleich,et al.  Rank aggregation via nuclear norm minimization , 2011, KDD.

[14]  Devavrat Shah,et al.  Rank Centrality: Ranking from Pairwise Comparisons , 2012, Oper. Res..

[15]  Emmanuel J. Candès,et al.  Exact Matrix Completion via Convex Optimization , 2008, Found. Comput. Math..

[16]  P. Moran.  On the method of paired comparisons. , 1947, Biometrika.

[17]  Arun Rajkumar,et al.  Ranking from Stochastic Pairwise Preferences: Recovering Condorcet Winners and Tournament Solution Sets at the Top , 2015, ICML.

[18]  Yuan Yao,et al.  Statistical ranking and combinatorial Hodge theory , 2008, Math. Program..

[19]  Martin J. Wainwright,et al.  Stochastically Transitive Models for Pairwise Comparisons: Statistical and Computational Issues , 2015, IEEE Transactions on Information Theory.

[20]  Inderjit S. Dhillon,et al.  Guaranteed Rank Minimization via Singular Value Projection , 2009, NIPS.

[21]  M. Kendall.  Further contributions to the theory of paired comparisons , 1955 .

[22]  Emmanuel J. Candès,et al.  The Power of Convex Relaxation: Near-Optimal Matrix Completion , 2009, IEEE Transactions on Information Theory.

[23]  Robert D. Nowak,et al.  Active Ranking using Pairwise Comparisons , 2011, NIPS.

[24]  A. Culyer.  Thurstone’s Law of Comparative Judgment , 2014 .

[25]  Andrea Montanari,et al.  Matrix Completion from Noisy Entries , 2009, J. Mach. Learn. Res..

[26]  Claire Mathieu,et al.  Electronic Colloquium on Computational Complexity, Report No. 144 (2006) How to rank with few errors A PTAS for Weighted Feedback Arc Set on Tournaments , 2006 .

[27]  James P. Keener,et al.  The Perron-Frobenius Theorem and the Ranking of Football Teams , 1993, SIAM Rev..

[28]  Mark Braverman,et al.  Noisy sorting without resampling , 2007, SODA '08.

[29]  Nebojsa Jojic,et al.  Efficient Ranking from Pairwise Comparisons , 2013, ICML.

[30]  T. Saaty,et al.  The Analytic Hierarchy Process , 1985 .

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

[32]  R. Duncan Luce,et al.  Individual Choice Behavior: A Theoretical Analysis , 1979 .