GLAD: Global-Local View Alignment and Background Debiasing for Unsupervised Video Domain Adaptation with Large Domain Gap.
Hyogun LeeKyungho BaeSeong Jong HaYumin KoGyeong-Moon ParkJinwoo ChoiPublished in: CoRR (2023)
Keyphrases
- domain adaptation
- manifold alignment
- semi supervised
- cross domain
- labeled data
- transfer learning
- target domain
- multiple sources
- pos tagging
- semi supervised learning
- supervised learning
- unlabeled data
- video sequences
- source domain
- test data
- sentiment classification
- concept space
- covariate shift
- unsupervised learning
- multiple views
- model selection
- active learning
- machine learning
- document classification
- facial expressions
- data points
- data sources
- reinforcement learning