MPAT: Building Robust Deep Neural Networks against Textual Adversarial Attacks.
Fangyuan ZhangHuichi ZhouShuangjiao LiHongtao WangPublished in: CoRR (2024)
Keyphrases
- neural network
- multi agent
- fuzzy logic
- digital image watermarking
- image watermarking
- natural language
- watermarking technique
- countermeasures
- recurrent neural networks
- neural nets
- semi fragile watermarking
- deep learning
- textual information
- fuzzy systems
- self organizing maps
- computationally efficient
- pattern recognition
- keywords
- metadata
- genetic algorithm