On the Complexity of Learning a Class Ratio from Unlabeled Data.
Benjamin FishLev ReyzinPublished in: J. Artif. Intell. Res. (2020)
Keyphrases
- unlabeled data
- positive examples
- active learning
- supervised learning
- labeled and unlabeled data
- learning algorithm
- labeled data
- semi supervised
- semi supervised learning
- co training
- class labels
- negative examples
- supervised learning algorithms
- learning process
- sample selection bias
- positive and unlabeled examples
- domain adaptation
- labeled examples
- learning problems
- learning tasks
- text classification
- training examples
- training data
- labeled instances
- multi view learning
- inductive inference
- training set
- data points
- prior knowledge
- active learner
- reinforcement learning
- labeled training data
- semi supervised classification
- high dimensional
- information extraction
- multi view
- class distribution
- learning models
- human experts
- unsupervised learning