Making Deep Neural Networks Robust to Label Noise: Cross-Training With a Novel Loss Function.
Zhen QinZhengwen ZhangYan LiJun GuoPublished in: IEEE Access (2019)
Keyphrases
- loss function
- label noise
- neural network
- stochastic gradient descent
- training process
- pairwise
- training algorithm
- support vector
- training data
- noise tolerant
- active learning
- noise model
- reproducing kernel hilbert space
- training set
- risk minimization
- image classification
- convex loss functions
- training examples
- markov random field
- boosting algorithms
- lower bound