Cross-domain Self-supervised Learning for Domain Adaptation with Few Source Labels.
Donghyun KimKuniaki SaitoTae-Hyun OhBryan A. PlummerStan SclaroffKate SaenkoPublished in: CoRR (2020)
Keyphrases
- domain adaptation
- cross domain
- target domain
- source domain
- multiple sources
- knowledge transfer
- covariate shift
- cross domain text classification
- transfer learning
- sentiment classification
- semi supervised
- source and target domains
- learning process
- labeled data
- reinforcement learning
- domain specific
- learning algorithm
- learning tasks
- text categorization
- active learning
- prior knowledge
- training data
- information retrieval
- test data
- multi label
- semi supervised learning
- text mining
- co training
- label information
- pairwise
- data analysis