Pseudo-labeled Bootstrapping and Multi-stage Transfer Learning for the Classification and Localization of Dysplasia in Barrett's Esophagus.
Joost van der PuttenJeroen de GroofFons van der SommenMaarten R. StruyvenbergSvitlana ZingerWouter L. CurversErik J. SchoonJacques J. BergmanPeter H. N. de WithPublished in: MLMI@MICCAI (2019)
Keyphrases
- feature space
- multistage
- transfer learning
- training set
- text classification
- labeled examples
- labeled data
- feature selection
- knowledge transfer
- machine learning
- supervised learning
- active learning
- single stage
- learning tasks
- semi supervised learning
- domain adaptation
- lot sizing
- dynamic programming
- manifold alignment
- transfer knowledge
- machine learning algorithms
- dimensionality reduction
- cross domain
- collaborative filtering
- multi task learning
- cross domain learning
- text categorization
- unlabeled data
- semi supervised
- learning algorithm
- document classification
- logistic regression
- optimal policy
- co training
- unsupervised learning
- domain specific
- support vector
- reinforcement learning