Investigating the Consistency of Uncertainty Sampling in Deep Active Learning.
Niklas PenzelChristian ReimersClemens-Alexander BrustJoachim DenzlerPublished in: GCPR (2021)
Keyphrases
- uncertainty sampling
- active learning
- cost sensitive
- experimental design
- random sampling
- query by committee
- learning algorithm
- maximum entropy
- semi supervised
- machine learning
- ensemble members
- misclassification costs
- learning strategies
- supervised learning
- training set
- generalization error
- learning process
- sampling algorithm
- training examples
- labeled data
- class imbalance
- data sets
- unlabeled data
- text classification
- support vector machine
- feature selection