Virtual Adversarial Training in Feature Space to Improve Unsupervised Video Domain Adaptation.
Artjoms GorpincenkoGeoffrey FrenchMichal MackiewiczPublished in: CoRR (2020)
Keyphrases
- domain adaptation
- feature space
- semi supervised
- training set
- supervised learning
- manifold alignment
- labeled data
- covariate shift
- training samples
- cross domain
- video sequences
- data sets
- semi supervised learning
- unsupervised learning
- high dimensional
- training examples
- training process
- pos tagging
- feature vectors
- multiple sources
- sentiment classification
- transfer knowledge
- kernel function
- training data
- feature selection
- target domain
- knowledge transfer
- active learning
- model selection
- text classification
- low dimensional
- data points