Multi-label active learning based on submodular functions.
Kuoliang WuDeng CaiXiaofei HePublished in: Neurocomputing (2018)
Keyphrases
- multi label
- active learning
- submodular functions
- graph cuts
- energy function
- greedy algorithm
- multi label classification
- text categorization
- image classification
- facility location problem
- multi instance
- binary classification
- multi label learning
- combinatorial optimization
- learning tasks
- image annotation
- text classification
- supervised learning
- class labels
- training examples
- objective function
- unlabeled data
- training set
- convex optimization
- learning algorithm
- learning process
- semi supervised
- labeled data
- learning problems
- class imbalance
- semi supervised learning
- machine learning
- cost sensitive
- generalization error
- markov random field
- transfer learning
- level set