Method and apparatus for training a translation disambiguation classifier

摘要

A method of training a classifier includes applying a first classifier to a first set of unlabeled data to form a first set of labeled data. The first classifier is able to assign data to classes in a first set of classes. A second classifier is applied to a second set of unlabeled data to from a second set of labeled data. The second classifier is able to assign data to classes in a second set of classes that is different from the first set of classes. The first and second sets of labeled data are used to retrain the first classifier.