Linking data separation, visual separation, and classifier performance using pseudo-labeling by contrastive learning.
Barbara Caroline BenatoAlexandre Xavier FalcãoAlexandru Cristian TeleaPublished in: CoRR (2023)
Keyphrases
- data sets
- statistical analysis
- raw data
- high quality
- synthetic data
- prior knowledge
- active learning
- data processing
- database
- original data
- background knowledge
- data analysis
- image data
- human experts
- data sources
- learning process
- data structure
- end users
- high dimensional data
- multimedia data
- reinforcement learning
- training data
- learning algorithm
- learning models
- knowledge discovery
- input data
- data collection
- data distribution
- learning tasks