QActor: Active Learning on Noisy Labels.
Taraneh YounesianZilong ZhaoAmirmasoud GhiassiRobert BirkeLydia Y. ChenPublished in: ACML (2021)
Keyphrases
- active learning
- label noise
- annotation effort
- training examples
- unlabeled examples
- labeling effort
- training set
- supervised learning
- learning strategies
- labeled examples
- learning algorithm
- semi supervised
- unlabeled instances
- labeled data
- active learning strategies
- relevance feedback
- selective sampling
- machine learning
- active learning framework
- class imbalance
- active learner
- neural network
- multi class
- pairwise
- learning process
- generalization error
- random sampling
- noisy data
- training data
- batch mode active learning
- unlabeled data
- data sets
- pool based active learning
- sample selection
- noise free
- experimental design
- multi label
- class labels
- missing data
- high dimensional