A Data Augmentation-Based Defense Method Against Adversarial Attacks in Neural Networks.
Yi ZengHan QiuGérard MemmiMeikang QiuPublished in: ICA3PP (2) (2020)
Keyphrases
- synthetic data
- input data
- neural network
- data sets
- test data
- prior knowledge
- noisy data
- data analysis
- data collection
- preprocessing
- training samples
- statistical methods
- significant improvement
- data sources
- information loss
- database
- spectral clustering
- data quality
- missing data
- clustering method
- correlation analysis
- image data
- knowledge discovery
- data points
- pairwise
- training data
- similarity measure
- data mining techniques
- high dimensional data
- artificial neural networks
- prior information
- feature space
- pattern recognition
- objective function
- neural nets
- training process