Detecting Errors and Estimating Accuracy on Unlabeled Data with Self-training Ensembles.
Jiefeng ChenFrederick LiuBesim AvciXi WuYingyu LiangSomesh JhaPublished in: NeurIPS (2021)
Keyphrases
- unlabeled data
- semi supervised learning
- co training
- semi supervised
- labeled data
- semi supervised classification
- training set
- ensemble learning
- supervised learning
- active learning
- labeled examples
- text classification
- labeled instances
- labeled training data
- labeled and unlabeled data
- training examples
- domain adaptation
- number of labeled examples
- prediction accuracy
- unsupervised learning
- label propagation
- data points
- abundant unlabeled data
- training data
- machine learning
- small set of labeled
- labeling effort
- text categorization
- decision trees
- learning models
- class labels
- supervised learning algorithms
- semi supervised learning algorithms
- learning algorithm
- instance selection
- single view
- multi view
- prior knowledge
- positive examples
- class distribution
- learning tasks
- transfer learning
- sample selection bias
- classification accuracy
- feature extraction