Improving Maximum Classifier Discrepancy by Considering Joint Distribution for Domain Adaptation.
Zehang LinZhenguo YangRunwei SituFeitao HuangJianming LvQing LiWenyin LiuPublished in: WISE (2) (2018)
Keyphrases
- domain adaptation
- joint distribution
- training data
- random variables
- marginal distributions
- target domain
- test data
- co training
- labeled data
- maximum entropy
- transfer learning
- semi supervised
- conditional probabilities
- cross domain
- multiple sources
- semi supervised learning
- class labels
- feature space
- feature selection
- unlabeled data
- document classification
- training and test data
- support vector machine
- feature set
- probability distribution
- sentiment classification
- learning algorithm
- training examples
- training samples
- graphical models
- support vector
- learning tasks
- training set
- natural language
- bayesian networks
- image segmentation