Semi-Supervised Empirical Risk Minimization: When can unlabeled data improve prediction.
Yuval OrenSaharon RossetPublished in: CoRR (2020)
Keyphrases
- semi supervised
- unlabeled data
- semi supervised learning
- labeled data
- empirical risk minimization
- active learning
- co training
- supervised learning
- labeled and unlabeled data
- semi supervised classification
- unsupervised learning
- labeled examples
- pairwise
- data points
- text classification
- prediction accuracy
- multi view
- training data
- prior knowledge
- training examples
- pairwise constraints
- supervised and semi supervised
- domain adaptation
- feature extraction
- learning algorithm
- machine learning
- transfer learning
- class labels
- training set