Active ranking from pairwise comparisons and when parametric assumptions do not help

DOI10.1214/18-AOS1772zbMATH Open1435.62293arXiv1606.08842MaRDI QIDQ2284367FDOQ2284367

Kannan Ramchandran, Nihar B. Shah, Reinhard Heckel, Martin J. Wainwright

Publication date: 15 January 2020

Published in: The Annals of Statistics (Search for Journal in Brave)

Abstract: We consider sequential or active ranking of a set of n items based on noisy pairwise comparisons. Items are ranked according to the probability that a given item beats a randomly chosen item, and ranking refers to partitioning the items into sets of pre-specified sizes according to their scores. This notion of ranking includes as special cases the identification of the top-k items and the total ordering of the items. We first analyze a sequential ranking algorithm that counts the number of comparisons won, and uses these counts to decide whether to stop, or to compare another pair of items, chosen based on confidence intervals specified by the data collected up to that point. We prove that this algorithm succeeds in recovering the ranking using a number of comparisons that is optimal up to logarithmic factors. This guarantee does not require any structural properties of the underlying pairwise probability matrix, unlike a significant body of past work on pairwise ranking based on parametric models such as the Thurstone or Bradley-Terry-Luce models. It has been a long-standing open question as to whether or not imposing these parametric assumptions allows for improved ranking algorithms. For stochastic comparison models, in which the pairwise probabilities are bounded away from zero, our second contribution is to resolve this issue by proving a lower bound for parametric models. This shows, perhaps surprisingly, that these popular parametric modeling choices offer at most logarithmic gains for stochastic comparisons.

Full work available at URL: https://arxiv.org/abs/1606.08842

Recommendations

• Simple, Robust and Optimal Ranking from Pairwise Comparisons
• A nearly instance optimal algorithm for top-\(k\) ranking under the multinomial logit model
• Rank Centrality: Ranking from Pairwise Comparisons
• Top-\(\kappa\) selection with pairwise comparisons
• An active learning algorithm for ranking from pairwise preferences with an almost optimal query complexity

zbMATH Keywords

rankingonline learningactive learningpairwise comparisonsmulti-armed banditsBradley-Terry-Luce model

Mathematics Subject Classification ID

Learning and adaptive systems in artificial intelligence (68T05) Paired and multiple comparisons; multiple testing (62J15)

Cites Work

• Title not available (Why is that?)
• On the complexity of best-arm identification in multi-armed bandit models
• Title not available (Why is that?)
• MM algorithms for generalized Bradley-Terry models.
• Title not available (Why is that?)
• Title not available (Why is that?)
• Regret Analysis of Stochastic and Nonstochastic Multi-armed Bandit Problems
• A Sequential Procedure for Selecting the Population with the Largest Mean from $k$ Normal Populations
• The \(K\)-armed dueling bandits problem
• Majorization, entropy and paired comparisons
• Stochastically Transitive Models for Pairwise Comparisons: Statistical and Computational Issues
• Simple, Robust and Optimal Ranking from Pairwise Comparisons
• Estimation from pairwise comparisons: sharp minimax bounds with topology dependence
• Competitive analysis of the top-K ranking problem
• Active ranking from pairwise comparisons and when parametric assumptions do not help

Cited In (14)

• Towards optimal estimation of bivariate isotonic matrices with unknown permutations
• Simple, Robust and Optimal Ranking from Pairwise Comparisons
• A General Pairwise Comparison Model for Extremely Sparse Networks
• Active ranking from pairwise comparisons and when parametric assumptions do not help
• Title not available (Why is that?)
• Title not available (Why is that?)
• Spectral method and regularized MLE are both optimal for top-\(K\) ranking
• Time-homogeneous top-K ranking using tensor decompositions
• Asymptotically Optimal Sequential Design for Rank Aggregation
• Pair-matching: link prediction with adaptive queries
• Low Permutation-rank Matrices: Structural Properties and Noisy Completion
• Optimal full ranking from pairwise comparisons
• Ranking and selection for pairwise comparison
• Robust Learning of Consumer Preferences

This page was built for publication: Active ranking from pairwise comparisons and when parametric assumptions do not help

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2284367)