Spectral Methods meet EM: A Provably Optimal Algorithm for Crowdsourcing

Yuchen Zhang, Xi Chen, Dengyong Zhou, and Michael I. Jordan

Abstract

Crowdsourcing is a popular paradigm for effectively collecting labels at low cost. The Dawid-Skene estimator has been widely used for inferring the true labels from the noisy labels provided by non-expert crowdsourcing workers. However, since the estimator maximizes a non-convex log-likelihood function, it is hard to theoretically justify its performance. In this paper, we propose a two-stage efficient algorithm for multi-class crowd labeling problems. The first stage uses the spectral method to obtain an initial estimate of parameters. Then the second stage refines the estimation by optimizing the objective function of the Dawid-Skene estimator via the EM algorithm. We show that our algorithm achieves the optimal convergence rate up to a logarithmic factor. We conduct extensive experiments on synthetic and real datasets. Experimental results demonstrate that the proposed algorithm is comparable to the most accurate empirical approach, while outperforming several other recently proposed methods.

Details

Publication typeTechReport
URLhttp://arxiv.org/abs/1406.3824
NumberMSR-TR-2014-116
> Publications > Spectral Methods meet EM: A Provably Optimal Algorithm for Crowdsourcing