One-to-N & N-to-One: Two Advanced Backdoor Attacks Against Deep Learning Models.
Mingfu XueCan HeJian WangWeiqiang LiuPublished in: IEEE Trans. Dependable Secur. Comput. (2022)
Keyphrases
- learning models
- machine learning
- loss function
- learning algorithm
- learning problems
- semi supervised learning
- learning paradigms
- classification models
- learning tasks
- conditional random fields
- machine learning algorithms
- natural language processing
- image segmentation
- data sets
- maximum likelihood
- higher order
- supervised learning
- probabilistic model
- machine learning models
- sparse metric learning