Active Learning for Efficient Audio Annotation and Classification with a Large Amount of Unlabeled Data.
Yu WangAna Elisa Méndez MéndezMark CartwrightJuan Pablo BelloPublished in: ICASSP (2019)
Keyphrases
- active learning
- unlabeled data
- labeled data
- supervised learning
- labeled and unlabeled data
- semi supervised learning
- co training
- semi supervised
- class labels
- selective sampling
- training set
- semi supervised classification
- improve the classification accuracy
- machine learning
- supervised learning algorithms
- labeled instances
- labeled examples
- label information
- text classification
- cost sensitive
- semi supervised learning algorithms
- learning algorithm
- labeling effort
- training examples
- semisupervised learning
- random sampling
- unlabeled examples
- labeled training data
- unlabeled instances
- classification accuracy
- sample selection
- transfer learning
- decision boundary
- class distribution
- unlabeled samples
- labeling process
- select relevant features
- text categorization
- unsupervised learning
- training data
- feature extraction
- random selection
- transductive inference
- support vector machine
- feature space
- learning process
- transductive learning
- training samples
- high dimensional
- feature selection
- supervised and semi supervised
- query by committee
- decision trees
- learning frameworks
- labeled data for training
- data mining
- support vector
- pairwise
- positive examples
- data points
- natural language processing
- image classification
- model selection
- learning tasks
- classification systems