Graph-based active learning: A new look at expected error minimization.
Kwang-Sung JunRobert D. NowakPublished in: GlobalSIP (2016)
Keyphrases
- active learning
- expected error
- training set
- semi supervised
- generalization error
- objective function
- supervised learning
- random sampling
- lower bound
- uniform distribution
- labeled data
- binary classification
- training examples
- uniform convergence
- semi supervised learning
- classification error
- worst case
- learning process
- linear classifiers
- learning algorithm
- machine learning
- cost sensitive
- neural network
- sample complexity
- belief state
- training error