Dynamic-Pix2Pix: Noise Injected cGAN for Modeling Input and Target Domain Joint Distributions with Limited Training Data.
Mohammadreza NaderiNader KarimiAli EmamiShahram ShiraniShadrokh SamaviPublished in: CoRR (2022)
Keyphrases
- training data
- target domain
- marginal distributions
- domain adaptation
- joint distribution
- labeled data
- source domain
- transfer learning
- cross domain
- input data
- unlabeled data
- test data
- sufficient training data
- labeled training data
- training samples
- semi supervised learning
- semi supervised
- training set
- data sets
- learning algorithm
- supervised learning
- training examples
- multiple sources
- maximum entropy
- domain specific
- machine learning
- conditional probabilities
- class labels
- classification accuracy
- similarity measure
- decision trees
- test set
- information retrieval
- training and test data
- transfer knowledge
- active learning
- random variables
- knowledge transfer
- support vector
- text classification
- data mining