Adversarial Crowdsourcing Through Robust Rank-One Matrix Completion

We consider the problem of reconstructing a rank-one 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.