Not All Unlabeled Data are Equal: Learning to Weight Data in Semi-supervised Learning.
Zhongzheng RenRaymond A. YehAlexander G. SchwingPublished in: NeurIPS (2020)
Keyphrases
- unlabeled data
- semi supervised learning
- labeled data
- labeled and unlabeled data
- semi supervised
- supervised learning
- co training
- active learning
- learning problems
- learning models
- training data
- labeled instances
- unsupervised learning
- learning algorithm
- labeled training data
- supervised learning algorithms
- prior knowledge
- data points
- partially labeled
- positive examples
- domain adaptation
- text classification
- labeled examples
- number of labeled examples
- semi supervised classification
- multi view learning
- machine learning
- improve the classification accuracy
- transfer learning
- training set
- regularization framework
- class labels
- unlabeled examples
- learning tasks
- label information
- semi supervised learning algorithms
- training examples
- data analysis
- reinforcement learning
- semi supervised learning methods
- label propagation
- data sets
- learning frameworks
- unlabeled training data
- semi supervised learning setting
- pairwise
- unlabeled samples
- decision trees
- unlabeled instances
- semi supervised clustering
- pairwise constraints
- sample selection bias
- human experts
- partially labeled data
- metric learning
- target domain