Certified Robustness Against Natural Language Attacks by Causal Intervention.
Haiteng ZhaoChang MaXinshuai DongAnh Tuan LuuZhi-Hong DengHanwang ZhangPublished in: CoRR (2022)
Keyphrases
- natural language
- digital image watermarking
- image watermarking scheme
- semi fragile watermarking
- natural language processing
- semantic analysis
- causal models
- semantic interpretation
- watermarking scheme
- malicious attacks
- causal reasoning
- watermark embedding
- malicious users
- question answering
- causal relationships
- countermeasures
- language processing
- neural network
- natural language generation
- geometric attacks
- natural language sentences
- dialogue system
- traffic analysis
- image watermarking
- natural language understanding
- wavelet domain
- computational efficiency
- knowledge representation
- recommender systems
- data mining