P-DIFF: Learning Classifier with Noisy Labels based on Probability Difference Distributions.
Wei HuQiHao ZhaoYangyu HuangFan ZhangPublished in: CoRR (2020)
Keyphrases
- learning algorithm
- learning tasks
- probability distribution
- supervised learning
- learning process
- learning systems
- learning phase
- machine learning
- training data
- learning problems
- active learning
- multi label
- label noise
- sample selection bias
- class labels
- unsupervised learning
- image classification
- training set
- reinforcement learning
- decision trees