In all LikelihoodS: How to Reliably Select Pseudo-Labeled Data for Self-Training in Semi-Supervised Learning.
Julian RodemannChristoph JansenGeorg SchollmeyerThomas AugustinPublished in: CoRR (2023)
Keyphrases
- semi supervised learning
- labeled data
- unlabeled data
- semi supervised
- abundant unlabeled data
- co training
- semi supervised classification
- training data
- active learning
- supervised learning
- text classification
- semi supervised learning methods
- labeled training data
- data points
- transfer learning
- semi supervised learning algorithms
- label propagation
- labeled and unlabeled data
- class labels
- learning algorithm
- labeled examples
- domain adaptation
- text categorization
- machine learning
- prior knowledge
- pairwise constraints
- number of labeled examples
- supervised methods
- unlabeled examples
- multiple instance learning
- data sets
- training examples
- maximum likelihood
- learning models
- unlabeled instances
- unsupervised learning
- pairwise
- training set
- target domain
- semisupervised learning