Text Laundering: Mitigating Malicious Features Through Knowledge Distillation of Large Foundation Models.
Yi JiangChenghui ShiOubo MaYouliang TianShouling JiPublished in: Inscrypt (2) (2023)
Keyphrases
- prior knowledge
- low level
- predictive power
- automatically extracted
- classification models
- knowledge acquisition
- co occurrence
- feature set
- general knowledge
- semantic information
- knowledge base
- semantic models
- feature vectors
- syntactic features
- additional features
- linguistic information
- knowledge based systems
- knowledge management
- database
- knowledge representation
- domain knowledge
- predictive features
- design process
- content features
- domain dependent
- structural features
- key concepts
- learning models
- statistical models
- expert systems
- feature space
- learning algorithm
- neural network