PseudoCal: A Source-Free Approach to Unsupervised Uncertainty Calibration in Domain Adaptation.
Dapeng HuJian LiangXinchao WangChuan-Sheng FooPublished in: CoRR (2023)
Keyphrases
- domain adaptation
- target domain
- multiple sources
- semi supervised
- source domain
- pos tagging
- manifold alignment
- labeled data
- cross domain
- transfer learning
- semi supervised learning
- unsupervised learning
- training data
- supervised learning
- test data
- unlabeled data
- domain specific
- source and target domains
- training samples
- covariate shift
- active learning
- databases
- co training
- sentiment classification
- data sets
- multi view
- document classification
- collaborative filtering
- data sources