SCAAT: Improving Neural Network Interpretability via Saliency Constrained Adaptive Adversarial Training.
Rui XuWenkang QinPeixiang HuangHao WangLin LuoPublished in: BMVC (2023)
Keyphrases
- neural network
- training algorithm
- training process
- feed forward neural networks
- multi layer perceptron
- recurrent networks
- feedforward neural networks
- neural network training
- training patterns
- training set
- artificial neural networks
- neural network is trained
- backpropagation algorithm
- train a neural network
- prediction accuracy
- neural network model
- feed forward
- online learning
- cmac neural network
- adaptive neural
- adaptive fuzzy
- neural network structure
- fuzzy neural network
- recurrent neural networks
- training samples
- supervised learning
- pattern recognition
- adaptive kalman filter
- neural fuzzy
- network architecture
- saliency detection
- adaptive control
- prediction model
- multilayer perceptron
- training examples
- object detection
- learning algorithm