Robust Assignment of Labels for Active Learning with Sparse and Noisy Annotations.
Daniel KaluzaAndrzej JanuszDominik SlezakPublished in: ECAI (2023)
Keyphrases
- active learning
- label noise
- ground truth labels
- semi supervised
- training set
- annotation effort
- training examples
- data sets
- noisy environments
- training data
- computationally efficient
- sparse data
- high dimensional
- metadata
- semantic annotation
- learning algorithm
- real time face tracking
- multi label
- erroneous data
- rank minimization
- label propagation
- compressive sensing
- cost sensitive
- learning process
- reinforcement learning
- machine learning