Online Active Learning for Cost Sensitive Domain Adaptation

Active learning and domain adaptation are both important tools for reducing labeling effort to learn a good supervised model in a target domain. In this paper, we investigate the problem of online active learning within a new active domain adaptation setting: there are insufficient labeled data in both source and target domains, but it is cheaper to query labels in the source domain than in the target domain. Given a total budget, we develop two costsensitive online active learning methods, a multi-view uncertainty-based method and a multi-view disagreement-based method, to query the most informative instances from the two domains, aiming to learn a good prediction model in the target domain. Empirical studies on the tasks of cross-domain sentiment classification of Amazon product reviews demonstrate the efficacy of the proposed methods on reducing labeling cost.