Disabling Backdoor and Identifying Poison Data by using Knowledge Distillation in Backdoor Attacks on Deep Neural Networks.
Kota YoshidaTakeshi FujinoPublished in: AISec@CCS (2020)
Keyphrases
- neural network
- data sets
- raw data
- data collection
- knowledge discovery
- data analysis
- synthetic data
- training data
- expert knowledge
- data sources
- pattern recognition
- domain knowledge
- database
- input data
- original data
- human experts
- data mining techniques
- end users
- prior knowledge
- missing data
- domain experts
- spatial data
- image data
- data quality
- enormous amounts
- hidden knowledge