Corpus ID: 59553474

Graph Resistance and Learning from Pairwise Comparisons

@inproceedings{Hendrickx2019GraphRA,
  title={Graph Resistance and Learning from Pairwise Comparisons},
  author={Julien M. Hendrickx and Alexander Olshevsky and Venkatesh Saligrama},
  booktitle={ICML},
  year={2019}
}
We consider the problem of learning the qualities of a collection of items by performing noisy comparisons among them. [...] Key Method We prove that, after a known transition period, the relevant graph-theoretic quantity is the square root of the resistance of the comparison graph. Specifically, we provide an algorithm that is minimax optimal. The algorithm has a relative error decay that scales with the square root of the graph resistance, and provide a matching lower bound (up to log factors). The…Expand
Minimax Rate for Learning From Pairwise Comparisons
We consider the problem of learning the qualities w1, . . . , wn of a collection of items by performing noisy comparisons among them. We assume there is a fixed “comparison graph” and everyExpand
Minimax Rate for Learning From Pairwise Comparisons in the BTL Model
TLDR
It is shown that the determination of the minimax rate is achieved by a simple algorithm based on weighted least squares, with weights determined from the empirical outcomes of the comparisons, which can be implemented in nearly linear time in the total number of comparisons. Expand
Ranking a Set of Objects: A Graph Based Least-Square Approach
TLDR
A class of non-adaptive ranking algorithms that rely on a least-squares optimization criterion for the estimation of qualities are proposed and shown to be asymptotically optimal. Expand
Rank Aggregation from Pairwise Comparisons in the Presence of Adversarial Corruptions
TLDR
This paper considers a setting where pairwise comparisons are initially generated according to a BTL model, but a fraction of these comparisons are corrupted adversarially prior to being reported to us, and provides a novel algorithm that provably filters out the adversarial corruption from data under reasonable conditions on data generation and corruption. Expand
The Performance of the MLE in the Bradley-Terry-Luce Model in $\ell_{\infty}$-Loss and under General Graph Topologies
The Bradley-Terry-Luce (BTL) model is a popular statistical approach for estimating the global ranking of a collection of items of interest using pairwise comparisons. To ensure accurate ranking, itExpand

References

SHOWING 1-10 OF 33 REFERENCES
Estimation from Pairwise Comparisons: Sharp Minimax Bounds with Topology Dependence
TLDR
This work considers parametric ordinal models for pairwise comparison data involving a latent vector w* e Rd that represents the "qualities" of the d items being compared; this class of models includes the two most widely used parametric models|the Bradley-Terry-Luce (BTL) and the Thurstone models. Expand
Worst-case vs Average-case Design for Estimation from Fixed Pairwise Comparisons
TLDR
It is demonstrated that consistent estimation is possible when the assignment of items to the topology is randomized, thus establishing a dichotomy between worst-case and average-case designs. Expand
Iterative ranking from pair-wise comparisons
TLDR
This paper proposes a novel iterative rank aggregation algorithm for discovering scores for objects from pairwise comparisons which performs as well as the Maximum Likelihood Estimator of the BTL model and outperforms a recently proposed algorithm by Ammar and Shah. Expand
Minimax-optimal Inference from Partial Rankings
TLDR
It is shown that even if one applies the mismatched maximum likelihood estimator that assumes independence (on pairwise comparisons that are now dependent due to rank-breaking), minimax optimal performance is still achieved up to a logarithmic factor. Expand
Rank Centrality: Ranking from Pairwise Comparisons
TLDR
Experimental evaluations on synthetic data sets generated according to the popular Bradley-Terry-Luce (BTL) model show that the proposed Rank Centrality algorithm performs as well as the maximum likelihood estimator for that model and outperforms other popular ranking algorithms. Expand
Learning with Limited Rounds of Adaptivity: Coin Tossing, Multi-Armed Bandits, and Ranking from Pairwise Comparisons
TLDR
This work derives essentially matching upper and lower bounds on the query complexity of r-round algorithms, and shows that Θ(log∗ n) rounds are both necessary and sufficient for achieving the optimal worst case query complexity for identifying the k most biased coins. Expand
A Statistical Convergence Perspective of Algorithms for Rank Aggregation from Pairwise Data
TLDR
This paper shows that, under a 'time-reversibility' or Bradley-Terry-Luce (BTL) condition on the distribution, the rank centrality (PageRank) and least squares (HodgeRank) algorithms both converge to an optimal ranking. Expand
Spectral MLE: Top-K Rank Aggregation from Pairwise Comparisons
TLDR
This paper describes the minimax limits on identifiability of top-$K$ ranked items, in the presence of random and non-adaptive sampling, and proposes a nearly linear-time ranking scheme, called Spectral MLE, that allows perfect top-K$ item identification under minimal sample complexity. Expand
The K-armed Dueling Bandits Problem
TLDR
A novel regret formulation is proposed in this setting, as well as an algorithm that achieves information-theoretically optimal regret bounds (up to a constant factor) is presented. Expand
Statistical ranking and combinatorial Hodge theory
TLDR
Hodge decomposition sheds light on whether a given dataset may be globally ranked in a meaningful way or if the data is inherently inconsistent and thus could not have any reasonable global ranking; in the latter case it provides information on the nature of the inconsistencies. Expand
...
1
2
3
4
...