Adversarial Crowdsourcing Through Robust RankOne Matrix Completion
Abstract
We consider the problem of reconstructing a rankone matrix from a revealed subset of its entries when some of the revealed entries are corrupted with perturbations that are unknown and can be arbitrarily large. It is not known which revealed entries are corrupted. We propose a new algorithm combining alternating minimization with extremevalue filtering and provide sufficient and necessary conditions to recover the original rankone matrix. In particular, we show that our proposed algorithm is optimal when the set of revealed entries is given by an ErdősRényi random graph. These results are then applied to the problem of classification from crowdsourced data under the assumption that while the majority of the workers are governed by the standard singlecoin DavidSkene model (i.e., they output the correct answer with a certain probability), some of the workers can deviate arbitrarily from this model. In particular, the "adversarial" workers could even make decisions designed to make the algorithm output an incorrect answer. Extensive experimental results show our algorithm for this problem, based on rankone matrix completion with perturbations, outperforms all other stateoftheart methods in such an adversarial scenario.
 Publication:

arXiv eprints
 Pub Date:
 October 2020
 arXiv:
 arXiv:2010.12181
 Bibcode:
 2020arXiv201012181M
 Keywords:

 Computer Science  Machine Learning
 EPrint:
 41 pages, 8 figures, NeurIPS 2020 to be published