SHIFT: A Synthetic Driving Dataset for Continuous Multi-Task Domain Adaptation.
Tao SunMattia SegùJanis PostelsYuxuan WangLuc Van GoolBernt SchieleFederico TombariFisher YuPublished in: CVPR (2022)
Keyphrases
- domain adaptation
- multi task
- transfer learning
- multi task learning
- learning tasks
- target domain
- transfer knowledge
- cross domain
- labeled data
- multiple sources
- knowledge transfer
- machine learning
- sentiment classification
- multi class
- semi supervised
- metric learning
- test data
- semi supervised learning
- reinforcement learning
- feature selection
- learning problems
- information gain
- unlabeled data
- multi view
- text mining
- machine learning algorithms
- text categorization
- data sets
- low dimensional
- dimensionality reduction
- probabilistic model
- active learning
- pairwise
- data mining
- databases