Multi-component transfer metric learning for handling unrelated source domain samples.
Chang'an YiYonghui XuHan YuYuguang YanYang LiuPublished in: Knowl. Based Syst. (2020)
Keyphrases
- metric learning
- multi component
- source domain
- transfer learning
- learning tasks
- target domain
- domain adaptation
- multi task
- semi supervised
- labeled data
- training samples
- semi supervised learning
- cross domain
- feature space
- knowledge transfer
- data distribution
- distance metric
- test data
- active learning
- pairwise
- machine learning
- training data
- reinforcement learning
- supervised learning
- learning problems
- training set
- data sets
- unlabeled data
- text classification
- text categorization
- dimensionality reduction
- distance function
- non stationary
- label information
- learning algorithm
- training examples
- knn
- labeled training data
- data points
- high dimensional
- multiple sources
- collaborative filtering
- unsupervised learning
- text mining
- feature selection
- labeled and unlabeled data