Training Robust Deep Neural Networks on Noisy Labels Using Adaptive Sample Selection With Disagreement.
Hiroshi TakedaSoh YoshidaMitsuji MuneyasuPublished in: IEEE Access (2021)
Keyphrases
- sample selection
- active learning
- label noise
- neural network
- training data
- training set
- training process
- training examples
- training algorithm
- noisy environments
- feedforward neural networks
- learning algorithm
- genetic algorithm
- supervised learning
- support vector machine
- data sets
- back propagation
- semi supervised
- fuzzy logic
- pairwise
- cost sensitive
- selection strategy
- pattern recognition
- machine learning
- deep architectures
- image processing