Multiclass Continuous Correspondence Learning
Brian Bue, David Ray Thompson · 2011
We extend the Structural Correspondence Learning (SCL) domain adaptation al-gorithm of Blitzer et al. [4] to the realm of continuous signals. Given a set of labeled examples belonging to a “source ” domain, we select a set of unlabeled examples in a related “target ” domain that play similar roles in both domains. We define a mapping into a common feature space using the pivot sample, which allows us to adapt a classifier trained on source examples to classify target exam-ples. We show that when between-class distances are relatively preserved across domains, we can automatically select target pivots to bring the domains into cor-respondence, allowing us to adapt a classifier trained on source data, to classify target data. 1 Structural Correspondence Learning for Continuous Spaces We extend the Structural Correspondence Learning (SCL) algorithm of Blitzer et al. [4] to the realm of continuous signals. SCL is a domain adaptation technique which creates a mapping between a “source ” domain consisting of labeled examples, and an unlabeled “target ” domain using a set of