A Novel Visual Interpretability for Deep Neural Networks by Optimizing Activation Maps with Perturbation.
Qing-Long ZhangLu RaoYubin YangPublished in: AAAI (2021)
Keyphrases
- neural network
- visual information
- visual features
- artificial neural networks
- color coded
- multi layer
- pattern recognition
- visual data
- feed forward
- real time
- information processing
- fuzzy logic
- fuzzy systems
- back propagation
- decision trees
- low level
- object recognition
- reinforcement learning
- computer vision
- genetic algorithm
- visual cues
- functional mri
- data sets