Exploiting Diversity of Unlabeled Data for Label-Efficient Semi-Supervised Active Learning.
Felix BuchertNassir NavabSeong Tae KimPublished in: ICPR (2022)
Keyphrases
- semi supervised
- unlabeled data
- active learning
- labeled data
- semi supervised learning
- unlabeled instances
- class labels
- label propagation
- labeled examples
- co training
- supervised learning
- label information
- semi supervised classification
- multi label learning
- labeling effort
- labeled and unlabeled data
- unlabeled examples
- text categorization
- pairwise
- training examples
- labeling process
- semi supervised learning algorithms
- unsupervised learning
- training set
- labeled training data
- machine learning
- learning algorithm
- partially labeled
- partially labeled data
- multiple instance learning
- training data
- fully supervised
- manifold regularization
- transfer learning
- domain adaptation
- unlabeled samples
- pairwise constraints
- selective sampling
- supervised methods
- multi view
- semi supervised learning methods
- positive examples
- semi supervised clustering
- weakly supervised
- learning frameworks
- text classification