There Are Many Consistent Explanations of Unlabeled Data: Why You Should Average.
Ben AthiwaratkunMarc FinziPavel IzmailovAndrew Gordon WilsonPublished in: ICLR (Poster) (2019)
Keyphrases
- unlabeled data
- labeled data
- semi supervised learning
- semi supervised
- active learning
- co training
- labeled examples
- training data
- labeled and unlabeled data
- text categorization
- data points
- semi supervised classification
- supervised learning
- text classification
- learning algorithm
- small set of labeled
- training set
- training examples
- supervised learning algorithms
- labeled training data
- unsupervised learning
- number of labeled examples
- labeled data for training
- semisupervised learning
- multi view learning
- machine learning
- select relevant features
- domain adaptation
- positive examples
- transfer learning
- class labels
- label propagation
- multi view
- prior knowledge