TextTricker: Loss-based and gradient-based adversarial attacks on text classification models.
Jincheng XuQingfeng DuPublished in: Eng. Appl. Artif. Intell. (2020)
Keyphrases
- classification models
- training data
- decision trees
- software quality classification
- data corruption
- feature selection
- multi agent
- learning models
- countermeasures
- database
- attribute selection
- watermarking scheme
- models built
- learned models
- feature subset
- feature set
- information retrieval
- data sets
- imbalanced data
- decision tree algorithm
- similarity measure